The discourse on artificial intelligence (AI) is riddled with breathless hype on the one hand and apocalyptic predictions on the other.
Journalists, civil society leaders, and public policymakers are all trusted messengers who shape our understanding of emergent and timely topics, like artificial intelligence. Yet many don’t know where to start. We need to develop a shared understanding of these headline-grabbing technologies so that everyone can jump into the conversation feeling informed, emboldened, and included.
On December 5, we brought together leaders from journalism, education, government, and philanthropy, to share how they talk about artificial intelligence in their work. The session covered smart practices for communication about AI both inside your organization and out in public. How can we navigate ambiguity amid changing technologies? What are common sources of confusion, and where can people go for clarity?
PANEL DISCUSSION FEATURING Victoria Houed, Alex Kotran & Clarence Wardell III
thanks so much for joining us today uhmy name is Vivian Schiller I’m theexecutive director of Aspen digitalwe’re a program of the Aspen Institutethat focuses on technology andInnovation and media and lately ai ai aiand its impact on just about everythingin the world so um I know some of you onthis call just having uh gone down thelist of people who are already online soas you know uh my background is as ajournalist um more than 30 years yearsum uh in often in executive uh uhleadership uh in news organizationsbefore joining the Aspen Institute butum once a journalist always a journalistthat’s who I am in my core and so wasthis project is particularly exciting tome we set out to help uh journalistsacross different Pats who were firstencountering artificial intelligenceafter uh the blast furnace known uh ayear ago known as the launch of chat GPTmany people who were not familiar withAI um were sort of struggling on whatthe stories were how to cover it so wewe did a bunch of work on that includingreleasing three different primers orprimers depending on what generation youcome from about artificial intelligencewe’re going to drop the links to thoseum in in the chat um but the funny thinghappened along the way um as more andmore people encountered our our primersand as we spoke to folks the feedback wewere getting was you know what it’s notjust journalists that need this um a lotof folks uh need help sort ofunderstanding Ai and understanding howto talk about Ai and so uh we want toserve those communities as well and moredirectly so our in our session todayyou’re going to hear from someabsolutely fantastic speakers experts ineducation um nonprofits government andof course uh from news media uh um sowe’re really excited about this programwe hope that um it will be useful to youwe particularly want to thank ourcolleagues at the seagull familyEndowment for supporting this work uhbefore I turn it over to our firstsession I just want to mention that thisis just one of many many many programsthat we’re doing in on AI if you havesigned up for this uh uh for this uhsession or if you’re maybe watching itafter the fact feel free to sign up forour newsletter um we will drop a link tothat in the chat as well um we’re doinga lot of work around the intersection ofartificial intelligence intelligence andlabor on the impact on elections anddemocracy on us and Global policy theintersection of AI and cyber securitybasically anything having to do with AIin the public interest uh which is whatmotivates us and gets us out of bed towork on every day so um with that thanksagain for being here I am now going tointroduce uh my colleague uh ElanorTurman who is our emerging Techresearcher and she’s going to kick usoff uh with our first panel Ellena overtoyou thank you Vivian um to kick thingsoff we are going to be hearing fromthree great panelists from governmentnonprofits and philanthropy about howthey navigate talking about AI in thecourse of their work so first I’d liketo introduce you all to Victoria howtVictoria is an engineer returned policyentrepreneur who has dedicated hercareer to improving the tech policyecosystem and the modernization ofgovernment in 2023 she joined the USDepartment of Commerce as the Directorof AI policy and strategy she also hasworked at Pop fox schit futures andadvised speaker Pelosi’s office on allthings Tech um next I’d like tointroduce you all to Alex KRON Alex havereceives strategy Partnershipsfundraising and external relations forthe AI education project prior tofounding the AI education project hebuilt the AI ethics and corporate socialresponsibility function for H5 Inc apioneering AI company in the legalservices sector prior to H5 he managedBrandon policy Communications for anumber of major companies and last butnot least I would like to introduce DrClarence Wardell III Clarence is achange maker committed to harnessing thepower of Technology data and evidence tocreate positive social impact andachieve more Equitable societal outcomeshe’s a senior program officer on theeconomic mobility and opportunity teamat the Villa Melinda Gates Foundationwhere he leads the team’s efforts toFoster Innovation to improve economicoutcomes for individuals in the UnitedStates with low incomes before joiningthe foundation Clarence served inseveral senior roles in the Biden HarrisAdministration and on the transitionteam um so those are our lovelypanelists and we will save 10 minutes atthe end of this section for Q&A soplease submit questions using the Q&Afeature um through throughout this eventand we will answer as many as we can attheend so to kick things off um I’d like tohear from each of our panelists abouthow AI is showing up in your work andwhat kinds of people you need tocommunicate with about AI um Clarencelet’s let’s start withyou sure um thanks so much Elanor um andthanks to the Aspen digital team forhaving us here today um really excitedfor this conversation um obviously verytimely um and will continue to be so umso maybe I’ll just say a little bit umyou know so both both in my current roleand and as you mentioned uh kind ofprior to to joining the foundation uhworked in the first two years of theBiden Harris Administration um primarilywith the domestic policy Council and theAmerican Rescue plan implementation teamum kind of in both cases kind of currentrole and I’ll get into that a little bitmore but certainly within theadministration um had an opportunity toum help lead uh keepy pieces of thepresident’s uh Equity agenda in work uhboth inclusive of his day one Equity AOum as well as some of the other threadsthere um and was really fortunate uhwhile working on the domestic policyCouncil which was the lead for theequity work um to help co-lead a processalong with our colleagues uh at theoffice of Science and Technology policyon uh equity and AI across the federalgovernment it was a process thatultimately led to the publication of thethe blueprint for the AI Bill of Rightsum and so in that case you know very umand this continues to be so I I think umyou know on one end of the spectrumright there’s there’s certainly kind ofa lot of um I would say like alarmistconversation around AI especially thefairly recent breakthroughs I mentionedearlier around uhGPT um and what it portends for thefuture whether that’s you knowdisplacement of work uh work in workersum as well as some more existentialthreats around what uh what folks areare concerned about um but in in thatcontext you know I think we also see andwe’re are very hopeful around theopportunities um that these tools couldcreate and I hope we you know have somespace and time to to to have aconversation about that I think there’sa lot of things kind of the near termwhere there’s um some really awesomeopportunities uh to maybe close some ofthe um uh economic divides um I thinkthat would depend on you know guardrails around use um how we um and youknow as a society how we decide todeploy experiment test um in ways thatare ethical and responsible um and soyou know that AI Bill of Rights uh whichwas kind of a precursor as folks know toum in many ways much more expanded and Ithink one of the longest executiveorders in the history of the countryit’s the AI executive order um butreally there really focused on in theconversation really centered aroundgreat there’s a lot of opportunity thatis held here how do we ensure againcreate the appropriate guard rails learnas we go and not repeat many of of themistakes that we have in the past whereI think you know we’ve we’ve seentechnology kind of be applied in waysthat um um you know limit opportunitiesyou know infringe on people’s civilrights Etc so that was really kind ofthe goal of the Bill of Rights engagedyou know we could get more into it youknow across and that was like um acrossthe government right you know everyagency lots of sub agencies you knowthere’s not an agency place in thefederal government you know that was notand will not be affected in the thestituents in which they work with thatwon’t be affected by these tools andTechnologies so that was one piece a lotof that has translated into thefoundation my role I sit on our economicmobility and opportunity team we makeinvestments that that try to improveboth the near-term economic security andlong-term Mobility outcomes for folks200% or below the federal poverty line Ilead our Innovation R&D and training andworkers supports portfolio so in a verysimilar way focused on again how are weyou know oftentimes what what happens isa lot of the Innovation happens you knowin the in the places where people havethe resources to pay for these toolsTechnologies kind of our end is like howdo we not again repeat that how do weshorten that lag how do we start todesign from the beginning and supportfolks that are working with folks in ourFocus population to understand how thesethings can be best applied for theirneeds right so how do we go from thatlike broadscale promise for a lot ofthese tools to actually making thatpractical and in practice um and andthat has to begin with the folks thatthat we we hope to deliver value forfrom from these tools and products soright um that’s a lot on my kind ofday-to-day and you know where where Itend to think about uh the promise of ofthese tools so you’re kind of workingnot only with the general communicatingto the general public about these toolsbut also have worked communicating todifferent agencies and kind of more atthe federal level um how agencies on thefederal side a lot of time spent withlike nonprofits now and organizationswho may not have the technical skill setbut know that there’s something here forthe folks that they that they work withamazing yes thank you um Victoria how isAI showing up in your work and who arethe types of folks that you arecommunicating about AI with yeahdefinitely so I am as mentioned workingat Department of Commerce I solely ormostly work with our chief data officeron um kind of thinking through our datastrategy as it relates to Ai and so I’mcurrently leading a working group on AIand open government data assets sobasically like we’re thinking throughwhat does it mean to have ai ready dataso when we’re thinking about like howpeople interact with a lot of largelanguage models like chbt and Bard umwhat kind of data are they actuallyseeing and can they be using all of ourvery accurate information for examplelike our Census Data our Noah data whensomeone asks like what are tell me aboutthe demographics of Washington DC umright now it’ll either give you superoutdated information or it’ll just tellyou to go to the census website and sowe’re trying to figure out like how dowe actually get our data to be not justmachine readable but machineunderstandable how do we make our datacrawlable um and just like again how dowe work with these bigger stakeholdersso kind of to your question of who wework with we’re literally working witheverybody from think tanks academicsstartups um researchers Stak like allthe stakeholders who kind of play intothis space and thinking about like whatwas our past open data strategy and howdo we kind of Meet the moment in AI umanother portion of my position issupporting our CTO and others in ourCommerce leadership and thinking throughour overall AI strategy like internallywithin Commerce so a little bit morerelated to the like om Mo that had comeout along with the AO so what does itmean to have like our own secure modelsinternally within Commerce to kind of dobetter discovery of our of the data thatwe sit on um that’s not open to thepublic uh what are our policies foremployees for using gen within uhDepartment of Commerce or how do wethink about using AI within our hiringprocesses and like who are we leavingout by by utilizing that and what arethe things we need to ensure when we’reusing those um that type of Technologyuh to not like leave out like the neurodversion and the people who don’t havelike very traditional backgrounds andthings like that um which has beenreally fun to work on because I was likeI’m very dedicated to like Talentgetting AI talent and or just like TechTalent into government and that’s beensomething I’ve working on for a longtime so thinking through how theseemerging technologies will kind of likeaffect that that line of work um veryfun and then just like at a very highlevel as everyone has seen the Commercehas lots of responsibilities in the AOthat came out and uh we all kind ofthere’s very few of us who are workingon AI within the department so we’re allkind of just helping each other navigatethat space and and get that long list ofthingsdone lovely thank you um Alex over toyou how is AI showing up in your workand who are the types of folks that youtalk to aboutAI yeah so I I’m a CE of a nonprofitthat that builds AI literacy curriculumso we’re spending most of our timeworking in schools with school districtLeaderswith um with tens of thousands ofteachers um and then as you as it’s bestdescribed is almost like the educationecosystem so if you think about all thefolks that are involved in or will beinvolved in helping the education systemuh evolve and whether it’s a pivot or aan evolution we’ll we’ll see but uh youknow basically step up to all thesethese new challenges that AI is going topose for the education system um youknow we build curriculum and so our ourtheory of AI educ is not a class calledintroduction to artificial intelligencein the same way that we don’t have aclass called introduction to Google it’sa horizontal technology that’s appliedacross all of uh the a student’s entirelearning experience um so we embedmodular curriculum that standards aligninto math science uh English and socisocial studies uh in addition tocomputer science um so education is inmany ways like the the front lines ofthis if you think about most Americanstheir interaction with chat GPT has beendriven by their kids are using it towrite essays whether or not that’scheating is I think a question that weneed to answer um it is going todramatically reshape the skills thatwe’re going to needum uh the University of Pennsylvania inopeni did some research uh in the springand they were looking at um what are thedrops that have uh significant exposureto large language models and if you lookat the list it’s it’s it’s a frighteninglist it’s it it includes Financialanalysts writers uh designers these arenot small mathematicians data scientiststhese are not like small Niche careersthese are really broad-based skill setsthat apply to dozens of careers youthink about what is someone in businessschool studying to do I mean it isfinancial analysis um and as we watchingthe capabilities of these tools evolveand and I think they’re really importantthing is this this idea that it’s notjust about chat gbt being good at youknow XY or Z task it’s the fact thatchat gbt can now use tools but when yougo and like if you look at my backgroundhere I used Dolly and J gbg4 um so notthree or four months ago there was anentire industry around promptengineering for uh AI art generators andpeople were publishing prompt books andselling them for hundreds of dollars anda lot of that is really out of date nowbecause you can have a semanticconversation with chat gbtit designs The Prompt it then goes anduses the AI tool um it can use webbrowsing tools it can um it can it’lleventually be able to use calculators umso you know I think where where we sitis in this kind of uncomfortable tensionwhere you know we don’t want to scarepeople and um we don’t know enough aboutwhether we’re going to hit some sort oftechnological roadblock so there’s likea lot of reasons why uh chat jbt orrather language models might notcontinue progress in a linear fashionwhether we run into technicallimitations in terms of compute or uhlimitations in terms of our ability toactually find high quality data um butso far what we’re finding is that a lotof those barriers that would suggestthat this is something that might bemore episodic and similar to differentpast hype Cycles um we’re not reachingthose yet and so I think we really needto come to terms with uh this reallyimportant conversation about um you knowthe way we talk about the future of workthe way we talk about career Pathwaysthe way we talk about what students evenneed to learn is going to have to changeuh the way we talk about computerscience is going to have to change andum so when when when I think about likecommunication it’s it’s it’s one of themost important components of even thework in education because there are allof these stakeholders and there’recumbent stakeholders and if we don’twant them to be resistant to uh thenecessary Evolution they’re going tohave to have we have to get very smartabout how do we talk to them about thisuh technology in a way that is inclusiveand makes them feel like part of thisproject as opposed to to Simply thefolks who are going to be disrupted umout of whatever uh position or or rolethey have right no that’s yes wonderfulI uh kind of building off of that um Iwonder since you work large with withEducators as you were saying um I’minterested in how those questions howwhat kind of questions are they askingyou and how have those questions changedespecially in the lastin the last year or so um kind of forall of you for all the folks that you’vebeen working with um Victoria andClarence and Alex like how have thequestions that how have the questionspeople been asking about AI how havethose changedrecently I think going along with whatAlex is saying I feel like I Iespecially as a person who’s kind ofworking on the side of innovation andembracing Innovation while alsodefinitely relying on people in ourdepartment to kind of figure outguidelines of like Risk Management andSafety and Security and things like thatI think I struggle with a lot of thejust like in in the media I thinkespecially like the Doomsday kind ofspeak and the fact that we are going tohave like issues with job loss and Etcbut as Alex was kind of saying I thinkthere needs to just be like anacknowledgement of the place that we’rein in like in like with the technologyand that it is now integrating in everyfacet of our lives and no matter whatwe’re going to just have to like kind ofcome to terms with where we are andeducate yeah like kind of how Alexsaying like educate young people on likehow to meet the moment and how tointegrate with these tools and um Ithink I kind of struggle with like a lotof the conversations being around howit’s going to destroy things but notwhich I think is very helpful withgetting things like the AO out andgetting people around um investing timeinto setting the S wordss and theFrameworks um but I also think it kindof gets people to struggle with theother sides of it like investing inInnovation research and development andlike really understanding like how weshould be kind of meeting the moment abit um so I think that’s been like kindof the the biggest difference is like aspeople are hearing more about it a lotof the conversation is mostlyaround the things that are going toreally suck about it which there aredefinitely potentially a lot of thingsthat will but yeah no for sure um uhClarence and Alex have you been alsolike meeting that kind of tension oflike the kind of very negativerisk focused which which is not as as asVictoria put very uh succinctly is notlike necessarily a bad thing but a badthing to focus on but how how do youapproach that that tension in uh whenyou’re communicating about AI in yourwork go aheadAlex Al so I think this is actually likein and and be Elanor you you’ll have toum you’ll have to forgive me for maybetrying to inject a little bit ofcontroversy into this into this panel umthose are the best panels right so I Iwould say this like when we’re talkingto teachers and we’re talking to umpeople who ultimately what we need to dois convince them that uh that that theyneed to be seeking out opportunities tolearn and Empower themselves uh with theskills and the tools the knowledge umbecause you know like the analogy thatI’ve been using a lot is you knowcomputers are like a bicycle for themind you know language models generativeAI these are supercars uh you can be thebest cyclist in your city you roll up tothe starting line with like Ripplingmuscles you’ve been training for yearsand below average student rolls up intheir Corvette and the fix and beforethe starting gun goes off you know who’sgoing to win that race and soirrespective of this question of likeshould we be using AI in schools there’ssome big problems I think there’s G betremendous challenges with learning lossuh it does get in the way of buildingfoundational skills as far as we’ve seenum unless tools are specificallydesigned to mitigate that and still theresearch is out um but we know thatyou’re going to have to have thoseskills so so that’s kind of the the interms of coms we have to find a way tostrike this positive note however Ireally am not comfortable with this youknow I’ve been in the AI space since2015 and you know from the moment Ijoined the space there’s this sort oflike Mantra of well there we got toharness the benefits in the upside whilemitigating the risks and jobs will becreated but or jobs will be destroyedyes but jobs will be created we don’teven know what jobs are going to becreated uh I think that’s a bit of a copout I think it is possible that yes thisis going to be similar to other umtechnology transitions but it could lookvery different and what we’re seeingright here is a a the the gamut ofskills that are at that are that thatthat language models can can replace arevery Broad and this may not be job lossbut it could be wage depression um youalso don’t need to like have 50% ofsoftware engineers get laid off for thisto be a crisis I mean it would be morelike 5 to 10% um and that and that alonewould be enough to uh Force you knowhundreds of thousands of students whoare thinking about going into computerscience fields to start to rethink whatthat means and if you’re not gettinginto a tier one school you may not haveuh good opportunities to get a job andthen the final thing that I think wereally need to be talking about isthere’s an NBR study I was looking atwhen uh when does capital investmenttend to concentrate in Automation and umand they were looking at like jobpolarization so like the erosion of themiddle class due to like offshoring andand mechanization of of factories andthey found that like something like 88%of uh that happens in the one in the 12months following a recession and so wehaven’t had a recession for 15ish yearsand so there’s this I think reallyimportant question of when the nextrecession happens whether it’s one twoor three years from now um companies aregoing to have these incredibly powerfultools at their disposal that are goingto enable them to significantly reducetheir headcount um and you know I I Iworry about there’s like that you don’twant to scare people off but you alsodon’t want to um um uhmisrepresent the challenge and and makethis feel like well this is all just awash like we can just sort of like youknow just just trust us trust thetechies we’ll figure this out like Ireally do think this is an all Hands-Ondeck moment and um I’m even coming tothis panel hoping to learn about whatdoes it look like to have a honestconversation that’s not necessarilypessimistic in doomsday but also not uhyeah anyways yeah that’s left there CL Idon’t know maybe you can bring us backfrom that really dark place no I don’tknow I mean you know this is um been ina couple conversations over the pastlike week and a half with some reallyyou know really smart folksum you know I think it’s all right and Ilike also like there’s there’s a lot ofquestions that are coming up I do thinkthere are folks that are like on the onehand scared by some of these things umother folks are really optimistic Ithink what is true is like there’s stillit’s it’s a lot of prognostication it’slike folks really don’twe don’t know like we are living in inlike this is you know I’m maybe classifyyou know I’m Elder Millennial um whichyou know I was like the first time Iheard that term I was a little takingback but um you know this is like thethird kind of like the really third likebig like Tech transition you know if youyou know that I’ve lived through in mylifetime which is kind of wild to say umand like the the second which I wouldlike just pegas maybe and others myquibble is like kind of Web 2.0 um wasreally where I was like more formativeyears and maybe able to like reallyremember like a lot of the things thatwe thought were going to be true aroundsome of those tools some of which cameto pass and then some a lot of things Ithink that like I could not haveforeseen personally right like just thecreation of influencer uh of like theseinfluencer roles right like what a biglike um you know like you see kids at apretty young age right who like want tobe these influencers and maybe in someways it’s just like you know a differentform of like actor or whatever but I itjust like could not you know foreseethat myself like 15 years ago was likewhat a big like industry that was it’sjust like one example I think we thoughtthese tools in many ways you know if yougo back to to here Square uh Iran GreenRevolution right like we thought thesetools would actually be really strongArbiters of like democracy rightelevating voices um and we’ve seen someof that I don’t know that like and youknow there were certainly some folks atthe time I you know I was pretty deep inthe space but not as like thoughtfularound kind of the misinformation andreally like tools of uh autocraticregimes uh which in some cases I thinkare winning out and how these tools areused and so I there’s just so much we wewe don’t know um you know I I you knowit was interest Alex brought up it wasat this dinner I was at earlier thisweek and we were all getting ready to goand you know just some guy you knowtossed out this really interestingquestion that got folks re-engaged andwas like I have two kids who arestarting college right now they ask mewhat should I major in what should Itell them you know and it’s you know inanyway you know we could bat that onearound a little bit but you know it wasjust so fascinating hearing thedifferent answers around the tablebecause just like this moment is like sodisruptive and I do you know I don’tknow you know that we know exactly wherewe’re headed but it is also you know Ithink maybe to Alex’s Point like it’snot something that is necessarily likewe are passive actors within like thisis a conversation in a moment like wecan help shape shape a lot of that um Iwould just say one thing I think thatlike you know and as Alex said you knowhe’s been in space since 20 2015 I willsay at the beginning of the BidenAdministration right likethe Dr Nelson alre Nelson who folks mayknow was running osdp came in with avision to do this AI Bill of Rights kindof the chat GPT conversation was notthis generative AI was not in the waterat that time like I was say like as oneof the folks who got a chance to kind oftouch that and work on it like we werethinking about kind of you know morepredictive you know kind of algorithmsand other automated systems not thesebreakthroughs which are now and and it’sinteresting because I do think like AIin some ways like when people say thatnow I think a lot of folks at least inand this is an interesting one for thejournalist here right kind of referenceor referencing really shend for for GPTor something like that now and I dothink it’s like it would be helpful forus to really hold a continue to hold amuch broader conversation and arecognition for folks that these toolshave been in our lives for a decade plusnow right like that are making decisionsand really what the Bill of Rights wastrying to say you know in many casesmaking decisions about what resourcesand things that we are able to accessright whether that’s loans at a bankwhether that’s housing you know thingslike that um that we have had notransparency around right like that uhwe’re not quite sure like what where thedata is being pulled from how thedecisions are being made and so that wasreally so I just want to say that likewhile there are questions around kind ofexistential concerns right there aresome near-term things and some thingsthat folks have been grappling with fora while that I don’t hope we don’t losesight of as well sorry Elenor I don’teven remember your original questionforet my screen there but uh hopefullyyou know uh is it interesting yeah nofor sureI ah you’re pulling me in two differentdirections one is like could shouldshould we kind of talk about potentialmirrors with like the rise of theinternet and stuff like that but I wantI want to kind of push against theinstinct to go really big and to getreally small and practical instead um uhat least for the purposes of this panelum I wonderwhatum I mean this has come up in youranswers oflike AI meant a different thing it’sit’s like automated decision systemright like the the the the understandingof um what this term is talking abouthas dramatically shifted in the pastyear because generative AI doesn’treally fit under the legal structure ofautomated decision system um Iwonder Victoria especially since youhave uh worked in what I’m sure is avery intense um environment of thespeaker’s office where you have to likeconstantly be getting up to date veryquickly on new tech issues I wonder umhow do you very practically go aboutvetting um sources or like where do youlook to for information on newtechnologies um in thepast uh how did that change in your inyour current position likehow when generative AI stuff startedcoming up where did where did you lookum for for information honestly it comesfrom it’s come from like so manydifferent places and I think there’slike the when I was first especially waswhen I was working in the speaker officegetting up to date on like just anyemerging technology um so at the time Ithink that was like the it was 2020 andthat was like the N was coming and doinga lot of um like discussions with thehill and um I think there was like acouple of really interesting likealgorithmic accountability bills likeshout out to be um work on that uh in intheir past life but I think that I uhfor me a lot of it’s like honestly justfinding people on Twitter finding likewho are like the key people who I knowknow a lot about that particular topicor like um oftentimes like journalistsplay such a big part for me just beingable to look through whatever if there’slike an article that’s like on AI Iusually go through and see who they’requoting and who like the most who theythink the most important people are umand it’s such a shame that yeah likeTwitter SLX is going in in into like areally interesting Direction and and Idon’t know what that even is going tolook like in the next three years um butthat has been like a really useful toolfor me but there’s been a lot of umotheremerging uh voices I feel like like I Ifollow last week in on AI or last weekin AI uh they do like a podcast which tobe honest like needs a little work butthey their written version of thepodcast that links to all the individualthings that are happening AI has beenreally really helpful um and I thinkpodcasts have been probably the biggestthing for me just because I am a personwho likes to like listen and do otherthings um and so usually when I’mactually talking to new people and orlike people who are just gettinginvolved in the technology and likewanting to learn new things I’ll sendthem like Ezra Klein’s podcasts on Alphafold so like the protein foldingstructures um used for like vaccine andDrug development curing diseases andstuff that podcast is like really goodintroduction to like the impact in likescience and research and it doesn’treally focus on the chat gbts and barsand the things that people are normallyseeing in the news but kind of like whatare the scientific impacts that we couldhave um and like where’s like kind oflike the positive direction that wecould be going uh so I think for me it’sit’s a little bit all over the place alot of it’s kind of based on people andand the fact that there are like tons ofexperts right now out in the ecosystemwho are in Industry on the ground umable to kind of like inform all of uspeople in the government on like theday-to-day of of how things are kind ofgoing no that’s awesome thank you forsharing specific resources um Alex orClarence do you have do you have um likewhere where do you turn when you orwhere what resources do you send tofolks um who are trying to get up todate on um this dramatically changingconversation um I I’ll be quick becauseI I see Alex has a lot of resources thathe dropped in the chat I mean um I thinksimilar to Victoria I mean I um a I feellike you’re asking me to expose myinformation diet which is not which isnot good only only what you’recomfortablesharing um you know I um big podcastperson same you know um and it is justso you know there’s just so much outthere right now it is honestly kind ofhard to narrow right like if you wantedto go really deep on kind of thetechnical aspects of it like there’s somuch content I think you know I am a bitmore on kind of the um you know theapplication side um I I would just plusone I think uh you know Ezra Klein who’slike his podcast is kind of high on mylike uh daily or weekly rotation youknow did a series of conversations uhincluding one with uh Dr Nelson who Imentioned earlier which I thought wasjust a really um a a challenging kind ofprovocative conversation I think in likehow to talk about this stuff withoutfear fear mongering and in a verypractical I think she has such a greatperspective on um you know how this isyou know some certainly in many waysthis is new new technology obviouslyright but we do have a lot of tools andFrameworks that are a bit liketechnology agnostic that to to beginfrom and we need to like patch repairwhat have you or whatever language youwant to use around kind of where futureproof what have you I don’t know but umI I just found that to be like a verykind of practical conversation uh andthen similar you know and as saw Alex alot of links in there um as much as I’mtrying to like wean myself off Twitterlike there’s not still not a betterplace for both sports like live sportscontent and then also um you know thethe unfortunate thing is like in thismoment I have seen a lot of the liketrusted voices that I have for yearsfollowed around a lot of these issueshave left that platform and um havedispersed other places so it is a bitmuch more fragmented ecosystem um thatuh in the moment that makes it a littlebit more challenging to kind of find TheTrusted voices and the folks that I’veyou know I followed for years sorryAlex I don’t know I mean honestly like Ithink this so so first of all you knowVictoria I think you nailed it um youknow Twitter I would add Reddit I wouldadd Discord um and and that’s not whereI go for trusted voices where I’m tryingto uh figure out how do I conceptualizeand talk about these Technologies in acoherent and cogent way forpeople um I would say like a lot of theresources that Clarence mentioned orthat Aspen puts together I mean that’sreally where we should be turning is andClarence to your point like this is notthe first time that we’ve had totranslate a complex technology to abroader audience and so I think learningfrom Best Practices there is is iscritical but uh what what Twitter andReddit and I I hate Reddit sucessful andI also have misgivings about Twitter butthe pace of these Technologies it’smoving so fast I mean like six monthsago I was I would be giving you know Iwas I was listening to podcasts andpeople were talking about you know inthree to five years we’re going to havetext to photorealistic animated videoand it’s going to replace TV shows andand stable diffusion and paa was justreleased this week and you know it’sactually one of the things in my one ofthe links that I shared and you youwatch this demo and it’s like oh we’rethere I mean it’s just happened um andyou know there’s there been other Ithink for me reddit’s important becauseyou it’s basically a bunch of nerdstrying to jail break chat GPT and try topush it to its limits and so it’s agreat way to actually understand some ofthe risk factors um I wouldn’t point mymom or a teacher there necessarily but Ithink for me being able to have acoherent conversation with an educatorabout like what do the risks actuallylook like I mean it really helps to toget away from you know uh uh likehypothetical theoretical Concepts tolike let me just give you an example ofwhat somebody did last week that I readum the other thing that I want to saythis is like and and I’m actuallycurious for the panels on this is like II think one of the biggest challengesthat we face in terms of communicationsis the fact that so I I read somewherethis was last week99.9% of people of chat GPT users arenot paying for plus so that meanseverybody’s 99.9% are using the freeversion which means they’re using GPT3.5 and my concern is that everybody’syou know GPT 3.5 is pretty cool but it’snot amazing it doesn’t replace I don’tI’m not able to use it to actuallyreplace real work gp4 which is what youget when you pay for Plus actually Imean I can write gr I do write grproposals and I don’t have to editsometimes um at all I copy and paste nowI spend time prompt engineering but thepoint is it’s much more powerful and sothe challenge we face is not you knowthe V like almost everybody virtuallyeverybody has not seen the powerfultools that we have seen and they arealso operating completely understandablythrough this uh uh environment whereSilicon Valley has misled us time andtime again I mean we heard about themetaverse we heard about nft we heardabout blockchain and crypto and I thinkfor a lot of people they are rightfullyasking the question is this really goingto be a conversation that we’re havingtwo or three years from now or is thisjust the flavor of the day um so I I Isee that to be like the foremostchallenge is convincing people that thisis you know and and and Clarence whenyou talked about that executive orderbeing one of the longest in historythat’s one of the things that Imentioned it’s like there was noexecutive order on the metaverse and andand so and so we already have somesignals that I think can help peoplesort of like digest that this is truly atechnology paradigm shift that probablylooks closer to like the internet or theright Brother’s first flight um but theytheir their spider their Spidey sensesare tingling and I I can understandwhy Alex I completely agree with you I Iwas saying the exact same thing as afriend like that CH JP like 3.5 is likezero comparison in my mind to four uhand and I think I think for a lot of thepeople that interact with now who kindof say what you say about like theydon’t they don’t like see the impactjust yet I feel like again like usingsuper concrete examples of of ways inwhich it’s already changed our world solike I I was listening to this professorkind of talking about like the theconcept of precision medicine and thefact that like in the past doctors aretypically they’ve just been like kind ofguessing based on their own experiencesand that’s like how they do a lot oftheir diagnosis for their patients butnow based on um like there’s no moreguessing anymore in medicine like a lotof the time they can use models tobasically optimize the pathway fordiagnosis and like therapeuticintervention and um and that’s like ait’s like a crazy concept to me thinkingabout like my past and like going to thedoctor and having it just be like basedon whatever doctor I get um that’s how Iwill be diagnosed and now it’s like somuch more based on the patient itselfand just using an example like that orthe what we’ve already been doing onclimate change like using satellite datato be able to predict like whenwildfires will start and how like whatthe magnitude will be things like thatum I think get people actually tothinking about what it will be like in 3to five years but for the people who arejust interacting with chatri t like thefree version I just like don’t thinkthey’re seeing where it can kind of takeus and and the kind of good and againthe bad of of what that might look likein the next three or five years but yeahare there otherum like you you mentioned like groundingthings in examples which I think is isis great because AI kind of by itselfdoesn’t really mean anything it’s justsuch a huge umbrella term um are thereother communication tips that you allwould give to um folks trying to likeboth learn about these Technologies andcommunicate about them in a way thatdoesn’t kind of fall into these traps oflike buying into industry hype whilealso being like grounded in how toolsare actually being usedtoday we have kind of like the examplike grounding things with examples umare there other like communication tipsthat you would that you would gravitatetowards honestly when I was telling myniece about this she was coming to visitme for Thanksgiving uh I was tell I waslike keeping up with all the open AIdrama on Twitter uh and honestly I feellike the the AI ecosystem is just so funlike it’s like it’s almost it’s it’salmost like watching like Congress wherethere’s like all these like differentplayers and um I feel like I also wasjust getting her really involved in theecosystem through like telling her aboutall the drama with like Sam mman andlike kind of making it about the peopleand who are building it and like likeMark Zuckerberg and Facebook and theiror meta and their kind of like um uhinfluences and like what’s driving themand Google and like their products likelike they’re building data Commons likethis public resource where um you cankind of like do in-depth visual analysison data which is super interesting butlike kind of like who the key playersare and and what’s like of interest tothem I think also kind of gets peoplemore a little bit more invested thanjust saying like here’s a cool tool umbut making it a little little bit moreabout like um the people and thepotential risks and making a little bitjuicier for the average person becauseobviously this stuff is super wonky whenyou get to the nuts and bolts ofit Alex do you have something toaddum yeah I think uh I think thethe Victoria this this idea of Chad gbtis like it’s doing a lot of our work forus right I mean like aied we used tospend a lot of our most of our time likeraising awareness about AI education andAI literacy and we take meetings withwhoever was willing to make time for usand every like what it even is AI nowthey they may not know anything about AIbut they at least are hearing about itand so if you think that is a huge assetfor us that we have like this opening toactually build knowledgewhereas before we had to create theopening um what’s worked best for me isjust like you have to show them youcan’t tell like you just have to likeand I and I a lot of my presentations Iwas just at the Gates Foundation umbefore here and I was running like aworkshop and the way I described thiswas like I was like imagine you know I’myour nerdy cousin and this is theholidays and we’re at the kitchen tableand I’m just like like I’m gonna imagineyou’re coming and sitting down next tome and I’m just going to show you abunch of cool stuff and I think peoplehave responded really really well insome cases viscerally to just seeingsome of the examples for themselves andthen most importantly give them thechance to build something that isculturally or contextually relevant totheir work or what they do um becauseonce someone sees how they can useCH or they can useour to createsomething that they find cool orinteresting or useful that’s that’s athat’s something that we could only havedreamed of two years a ago you know andwe were thinking about how do you teachkids about artificial intelligence Imean we it literally our version of thatwas a nwe course where they had tobasically uh train a logistic regressionmodel it was very powerful asset I thinkthere are responsibilities involved andyou you need to be give them someguidance so that they’re able to use itsafely but um but yeah I mean justgetting handson islike it’s both yeah that’s that’s morekey pieces yeah no totally that thatmakes sense they’ve got like kind ofgoing fromhaving people actually interact with thetools that we’re talking about to likegrounded in their ownexperiences um we we talkingabout sticking with examples um examplesof how those tools are being used byother people out in the in the world andum as Victoria mentioned uh highlightingthe people like the actors that areusing these tools like there isn’t justan AI out there that’s kindof sensient and nebulous and like goingto solve all your problems it’s thereall these are created by people rightand there is drama and uh goofinessassociated with that um I’m going tostart to Pivot us into your lovely uhaudienceQ&A um and just for uh anyone everyonewatching along at home um we willcompile all the resources that have beenshared so far and send it out as afollow-up email so don’t feel like youhave to like desperately copypastethings from the chat um well we got youum so pivoting to audiencequestions um when talking about AI howcan we use language that empowers peopleto avoid being passive observers andassuming kind of this like negativeeverything’s inevitable like this isjust we we I’m this is not the way theperson phrased it but like justsubmitting to like the kind of negativeuh Silicon Valley kind of how how howhow would you use language that empowerspeople and makes them makes themactive I mean I just I don’t know that Ihave a great answer around language Imean I think so much of this comes backto you know the thing that in Victoriaand Alex mentioned I think the you knowthe way to empower folks is to um youknow bring it down from the you know30,000 foot level and it’s really likethose tangible Hands-On examples I meanI think that’s like the beauty of of achat GPT or something like that right ityou can viscerally you know know getsome like oh hey here’s what this isright like go ahead ask it a questionright like now now let’s like and so Ithink for me that’s that’s the piece ina way that um you know as Alex wasmentioning before you know somethinglike blockchain or metaverse you knowthat was like it was very abstract therewas a certain like there was a highthreshold to even begin to like engagewith something like that so it feelsvery far off um in a way that I thinkyou can kind of personalize umpersonalize the you know the experiencethat folks have with with these tools umas one way um I don’t know maybe othershave a better example around aroundlanguage but I do think it’s you knowsome of those it’s like a getting folksexperience and then you know having someexamples right where where where folkshave already you know experienceexperienced it and maybe maybe hopefullyhad a good experience with it um uh so Idon’t I always found blockchain stuffextremely confusing I will just say saythat about myself um and didn’t didn’treally feel like I could talk to peopleabout it because I felt very silly notunderstanding it um uh Alex and Victoriado you have tips on how to use how tocommunicate with folks about AI in a waythat empowers them to like ask thosesilly questions and feel like theybelong at the conversation like belongat the table having theconversation I think as Alex kind ofalluded to I think it really is justmeeting people where they are and kindof relating it to their particularinterests and things that they liketheir industry and their ecosystem andtheir life um because usually like I’mnot I don’t usually use like the exampleof like climate change or medicine orsomething when I’m talking to my dad whoworks like in bill collecting um and soI I feel like I it’s it’s really aboutcreating a safe space as we call it youknow of of using language that umdoesn’tI don’t know scare people away and alsomakes them feel very comfortable justasking silly questions and also notgetting mad if someone says AI insteadof machine learning which I kind ofreally those types of people I’m likeyou’re not making this a walking me likeexperience or whatever uh for people andobviously we’re all learning and and aswe talked about earlier a lot of thelanguage is changing over time um so andwe are kind of like through creators ofour own reality and so if we want to usecertain terminology for certain thingslike that is okay uh and peopleshouldn’t be kind of like shamed forthose specific things as long as theyhave like an understanding of um likewhat they need for to live their livesum but yeah I think a lot of it’s justkind of creating that that safe spacefor people uh and allowing them to beable to kind of ask those sillyquestions and not teing people apart onTwitter if they use the wrong um youknow Silly thing in their tweet soLife’s TooShort Alex yeah I usually I mean I I I Iusually whenever I’m doing let’s saydemos or presentations I’ll often try touse a tool that I haven’t played withuntil you one case I did a presentationwhere I had just it I was just usinganthropic to for the first time like 15minutes before the presentation and Idid that on purpose because I I I I Ithink it’s important for people tounderstand that everybody is basicallyat the beginning of their LearningJourney I mean maybe I’ve had a couplemonths you know Head Start but you knownobody’s an expert on language modelsand even the people who build thelanguage models you know they theythemselves can’t explain exactly how youknow i’ I’ve had the I’ve done a um youknow opening I did a a workshop with ourorganization back when they they firstgave us a grant and we were expectinglike you know a really deep dive likehow we’re g to come become power usersand and basically they were just likeyou know what it kind of comes down toexperimentation like you just need to BUlike like get comfortable with tryingsomething and and the best way to learnis just to experiment the the the secretthe secret sauce is experimentation plusdomain expertise so find an area thatyou know well it doesn’t matter what itis whether it’s cooking or sports orwhatever it is and then start using AIin that domain because you’re going tobe best equipped a to identify theissues of hallucinations but you’re alsogoing to figure out like there are someuse cases that just aren’t great usecases and there are some really amazinguse cases and um any attempt create LikeAplaybook would be rendered out of datetoo quickly so that’s a challenge from afrom a learning perspective becauseobviously curriculum is going to have tobe updated in real time but I think italso creates this opening becauseanybody can really have Entre to thislike this is there is no technicalbarrier anymore in terms of using thetools if you want to build an AI modelwell that’s a different story but veryfew people will be actually building themodels um even fine tuning models youcan we just did this hackathon it’s allsemantic language you just sort of likeprovide colloquial instructions to um tothe rapper and um so so that that is theup that is theglassical perspective on all of this isyes it’s like really complicated it’smoving fast but truly there there thebarriers that we face with ComputerScience Education uh don’t exist anymoreum there are challenges but there are Ithink significant opportunities and ifif we think about that and through thelens of equity and really focus on thatI think that there could be a ofupside for sure awesome thank you um I’mgoing to Pivot to another question fromthe audience um how can we talk about AIin ways that consider differentsociocultural contexts rather thanhaving only an American Centric SiliconValley Centricview does that come up in any of your inany of yourday-to-day umyou know I mean I guess it comes up inmy day-to-day in the sense that I I Ispoke about before you know I spend alot of time with nonprofit organizationsmany of which are working with folks whoI mentioned before are uh you know aboutyou know below that kind of 200% uhbelow the federal property line right sofolks from a um you know differentsocioeconomic um background who need umyou know who’s maybe near-term needs aredifferent um whose engagement with uhthese types of toolsum um you know I’m not going to say isit well I would say maybe limited or orisn’t it comes from a differentperspective right so um you know I Ithink the like you know to me so much ofthis has to be rooted in like what whatproblems are you solving for folks andnot not just like me certainly but likethe organizations that are working withwith folks like and and this is truerightlike this isonly any of this only means somethingright if you’re delivering value forsomeone right and like value isdifferent depending on where you whatyour what your background is whereyou’re situated in the world uh whattypes of resources you currently haveaccess to what your goals andaspirations are right and so I think youknow for me it’s like getting proximateto those folks getting approximate tothe to to folks who maybe have adifferent set of questions that theywant to ask want to ask and wantanswered by these tools um to begin toexpand um what they can do how weexperiment Etc and like and so that tome comes back a bit to what I talkedabout before I mean so much ofhistorically you know what has been thelead and what you know will probablycontinue to be the lead right is like umwe end up solving the problems for thefolks that have the resources that arethe F first movers right um what we’retrying to do here right is to bring someof that to folks who maybe don’t aren’tnecessarily part of these conversconversations and and don’t always havethe resources um or cultural SocialCapital to be part of thoseconversations right and so how do wehelp bridge that in some ways and inthis moment so um I don’t have like ablanket answer but I guess my like kindof like AI agnostic answer is um youknow a understanding what problems andchallenges the folks are dealing withand how do you make the tool whatever itis Tech or not relevant tothem no that totally makes sense it’s umthat seems that that has been like atheme of this entire conversation rightlike or it has felt that way to me ofjust rootingall of these conversations about verytechnical topicsin uh examples or things that actuallymean something to the people that you’retrying to communicate withum I since we are approaching one Iwould like to give all of our panelistsum time to for any last thoughtsanything that you would like to kind ofclose this out with um no no pressureobviously but um any burning thoughtssomething you wanted to say but couldn’tsay um because of the flow of theconversation all right Ium yeah thank you all I will end us alittle I will end us a couple of minutesearly thank you all who uh folks who arewatching for your wonderful questionsI’m sorry we didn’t get to too many ofthem um thank you to our wonderfulpanelists for a great and like wide-ranging discussion that still felt likewe were really uh I don’t know groundedtalking about like how how how to getthings done um thank you for sharing allof your resources and again we’ll sharethose out with um with attendees in anemail followup to this event um so thankyou Victoria thank you Alex and thankyou Clarence for your time um andexpertise and uh yes thank you everyoneuh for joining us for thispanel
Fireside Chat with Sisi Wei
really grateful to everyone for joiningus um and excited to start thisconversation so uh I’m gonna try to uhnot uh geek out too much here but I amso excited to be chatting with CCAeditor-in chief of the markup the markupis a nonprofit investigative Newsroomthat uh challenges technology to servethe public good we’ll dig into that alittle bit um before joining the markupCC was co-executive director of opennews she also served as assistantmanaging editor at propublica from 2018to 2020 where she managed a large interlarge interdisciplinary investigationsacross newsrooms one of which w a peitSur prize for National reporting in 2020CeCe also serves on the board of thenews Revenue Hub and the open notebookshe is so cool I’m gonna try to be cooluh really excited for this conversationum CeCe thank you so much for joining ushere um and before we get started hereuh I I I was doing a little a little bitof research and uh I I wanted to spend aminute here let’s just let’s answer thequestion how you feeling how you feelingoh how am I feeling uh what a moment intime it is uh I’m really excited to behere by the way thank you for doing thisbe and everyone at ask been um I feellike uh you know December is the time ofyear in journalism where we publishendlessly more so than any other time ofthe yearand um you know I think we are here totalk about AI which will continue tohave great impact on society andsimultaneously with that the you knowthe world is going through a lot rightnow we have two active Wars going on uhand um it seems like every day there’ssort of a new twist in what people arereally concerned about because there’s alot of concerning things outthere absolutely absolutely yeah it’sit’s an overwhelming time um in so manyrespects and I feel like that inparticular means that journalism and umplaces we can go for information areonly uh the more valuable and the morepowerful and um I I feel like your rolein particular as someone who acts as aneditor is really interesting because youknow you’ve been a leader not just inhelping the public to understand thingsbut also as an editor you know you’veyou and through a lot of the work inyour career you’ve done a lot of work inhelping journalists to understand thechaotic and confusing World um that welive in and and uh particularly throughthe lens of technology so I wonder ifyou could open us out here talking alittle bit um like Elanor asked in ourpreviouspanel what are some of the the peoplethat you find yourself talking to whatare the categories of folks that youfind yourself talking to and what areyour recommendations to folks as they’retrying to wrap their heads aroundAI yeah that is such a great question umand I think there’s sort of two elementsto this as a journalist we’re oftentrying to understand exactly to the mostgranular detail possible how somethingworks and um I’m sure we’ll talk aboutthis more later in terms of like how dowe communicate about AI because it’salso important to really understand itat a deeper level but in terms of thepeople that I’m talking to or even let’sjust broaden that out to the types ofpeople that the markup is talking towhen we do all of our stories right wehave a mix of people who are always theactive users of whatever technologywe’re reporting on or whatever algorithmor AI that we are writing about and thenalso the people who are directlyimpacted by it um in a lot of our worksometimes people um how do I phrase thisthe a really common experienceregardless if it’s about AI or any otherkind of tech is that when it impacts acommunity they often can feel theeffects and know it’s happening but theydon’t know that it’s happening literallyas a result of a specific technology orsomeone using that technology right sowhen we talk to them we often are end upbreaking the news to them exactly whythis feeling or this consequence thatthey already know might be happening andthen once they know we follow up afterthe fact as well um um and it is sort ofthe cycle of knowing both knowing thatcommunities sort of know when things aregoing on even if they don’t knowliterally how it came to be or if theyonly know some of the elements right butthey don’t know the specific Techelement if we can add to that knowledgethat’s also one of the things where umlike communicating can actually equippeople to take more action to have moreagency um but on the second part of yourquestion right where can people go ifthey’re just trying to wrap their headsaround AI um I think the panelists gavegreat answers so I will give a very verychill answer which is if you just if ifyou’re like AI isoverwhelming I don’t know what to dowhich a lot of people feel that way andyou kind of want to just get some likefun examples I actually highly recommendthat you go to Tik Tok go to Tik Tok andsearch for like Ai and the amazing thingabout Tik Tok is that the videos aregoing to be like 10 seconds or 30seconds long and they’re highly focusedon actually showing you how to do thingslike you’ll get really funny uses whichI think is really casual and inviting topeople but then you also really getyou’ll get some serious work uses aswell and through that you might actuallybe able to connect to it in a way thatdoesn’t feel so highstakes great call out and yeahabsolutely I think there’s such a aworld of possibilities and and evenknowing where to go um uh within thatworld of possibilities can feel like alot so having specific examples likethat is really great what to look upwhat terms to use love to see it um II’d love to dig in a little bit morethough you know the markup speaks to anaudience of people but you CC way speakto an audience of people as an editorand I’m curious if you can talk a littlebit about like how you uh how youapproach conversations with otherjournalists about this topic thesedays yeah um you know with otherjournalists this is something that I’vebeen trying to figure out for maybealmost a year now as well which is thatwe work with local journalists all thetime and also other National journalistsand one of the things I’ve been noticingis that similar to what I said aboutcommunities journalists know whenthere’s a problem uh and they sometimesknow like I just have to investigatethis person this organization thisprocess and then I’ll be able to to getto the heart of the problem butsometimes there’s a technology barrierwhere they’re like ah if only I couldget access to every example of this inthe city right like every parking ticketin the city and the reason why thenmaybe I could do more with it and therehave been moments in which people havecome to us at the markup being like canwe get some collaborative help to helpus sort of unlock information um and alot of it is around algorithms in someway which to me is like the theuh theoriginal Gem of where AI is coming fromright there’s algorithms in anythingthat has to do with AI um and what Ioften say to people is one that you’renot alone right if you’re confused aboutit if you don’t know how to use it ifyou don’t know how to use it in yourwork as a journalist you’re not alonejust everyone and every industry istrying to figure it out um and then theother thing is that we did this surveywith a lot of local partners that we’reinterested in asking them if we couldhelp you cover AI better right tell usmore about what your needs are and therewas one question where we asked peopleto just be really honest and we gavethem some pretty sassy opinionatedchoices where they could pick everythingthat resonated with them and literallythe most popular survey answered to oneof our questions um where people feltlike figuring out how our Newsroom willcover AI feels important but too hard totackle that was the most popular surveyanswer right um and I you know that issomething that resonates with me a lotbecause a lot of my work previous tobeing an editor was teaching journalistshow to code right because that openstons of doors for what you can do um butI’ll say one second thing a piece ofadvice I give to journalists is that nomatter what you cover also don’t ruleout ai’s relevance right some peoplelike oh I cover food I cover Arts Icover you know um um whatever topic andthey don’t think AI you know issomething that they have to payattention to but actually I think it’sgoing to permeate right it has beenpermeating already into everything umand I promise that it will becomerelevant and so the more that you atleast play with it now the better itwill allow you to be able to report onit when you really really have directinteraction with it um and then I dohave one last thing and I think thisgoes for everyone which is I give thisadvice to new reporterwho are covering technology but itapplies to journalists it applies toEveryday People right and it’s that ifyou are writing about AI or trying tounderstand it and you just don’tunderstand how something works you mostlikely will end up asking questions ofsomeone who you think does understandhow it works right so in journalism wedo this through interviews but anyonecan do this with someone that they haveaccess to or want to call up and if theperson explaining the technology to yousays something that you can’t understandask them to explain it another way neverbe intimidated by the technology becauseif someone can’t explain it to you in away that you get either they’re notgreat at explaining the topic whichvalid we’re not all great at explainingeverything or two they don’t actuallyfully understand it themselves andyou’re helping them realize that but itis never number three you individuallydon’t have the capacity to understand itthat is never the case right so if youdon’t get it just keep acting um andthat’s how you’re actually really ableto get at the root of how somethingWorks um never doubt yourself and yourcapability andunderstanding absolutely absolutely lovethat and I think that you know the inthe work that you do you have had somany you know throughout your career somany resources for other journalists somany resources um for folks who arelearning about coding learning about atype of um data reporting um and andthat kind of intellectual humility butalso a little bit of like no I deserveto be able to know this like I I shouldbe able to follow up um I think it’sreally powerful and I think um folks inin the journalism field uh famous foryour questions uh so really interestingto be on the other end of it asking youquestions here and I’m I’m actuallycurious you know sometimes it’s helpfulfor folks to have specific examples andI’m putting you on the spot here alittle bit but are there particularquestions that you find to be really uhyou know expansive or there questionsthat you wish more journalists would askwhen talking to people about theseTechnologies yeah I would say there’sprobably two buckets of questions um thefirst bucket gets into what I was justtalking about which is dig into how thealgorithm or how the AI works and what Imean by that is sometimes um and youknow our profession is responsible forthis too where you read about somethingand you know it’s AI or machine learningor algorithms because there’s onesentence that says like you know so andso organization use machine learning todo this and then it just tells you aboutwhat they do but not about the tech andI would say never let that go always asklike okay but what does that mean whatdo you actually use it for um how doesthat actually play into everything elseyou’re saying this is doing is itbecause you have a database of people’sspaces and the technology is facialrecognition uh do you have access todescriptions of a ton of famous art andthat’s how your you know like artgallery AI Works um how exactly does itwork because sometimes you’ll discoverin the previous panel I think got atthis as well sometimes you’ll discoversomething is actually uh not really AIuh it’s it’s something else uh that youcan actually very ex easily explain toother people but because it’s hiddenbehind this term that that could meananything right there’s no Clarity andthen a second category of question isreally asking questions that lead towhat is the impact of this AI right soright now ai is emerging right um inlike every technology before it it hasabsorbed all of mankind humankind’sproblems like racism sexism Etc becausethat is all the information it’s hadaccess to to train up right and so ofcourse theoretically it can be used in away to purposely show that bias likethat is not surprising to me when anyarticle like that comes out it’s notsurprising to the public uh but the keyto make it really interesting is thatthat’s only 50% of the story right theother half is trying to figure outasking questions so you can narrow downwho’s actually using it that way versustheoretically to prove a point and thenalso if you figure out there are peopleusing it that way or some way similarwhat then is the consequence like whathappens as a result of how they’re usingit are they using it to make decisionsand who is subject to those decisdecisions right and so those two bucketsI think are the main things when youwant to know more about a specific AI oralgorithm or any piece of tech honestlywhat I wouldask fantastic yeah I feel like that thatpoint you made about um you know thesort of aisc of using a big blanket termlike AI is something that’s reallyresonant and and it’s not alwaysnefarious right it can be someone’s ownyou know unfamiliarity or ignorancewhere like they just use the bucket termuh to explain everything but um but thatlevel of specificity and digging in isso powerful that being said it is alsothe case that you know I I I fancymyself a a metaphor Smith of sorts umthere is power in using metaphor simileanalogies uh other parts of speech thatI won’t be able to remember um right andand in your work you have used thesetools you’ve used games to communicatejournalistically you’ve used these toolsso I’m curious when you’re thinkingabout um communicating about technologyAI in particular but even technologymore broadly how do you kind of balanceusing you know evocative descriptionsthat help people to place a mental modelof something against the specificity andaccuracy of what’s actually going onbehind the scenes yeah um the solution Ithink is both extremely simpleuh but extremely complicated which isyou definitely use both right it’s kindof like I’ll draw a different parallelI’ll use a simile to explain our metadiscussion about similes um you know ina lot of the journalism we do at themarkup it’s highly technical right we doa lot of data analysis or we’re tryingto sort of look at an algorithm and thentell you how it does or doesn’t work anduh we can’t write an article that onlydoes that right we also have to sayexplain to you a human being andintroduce you to them and then show youlike how this interacts with theiractual life and why it matters and it’sthe marriage of those two things thatmakes journalism really powerful rightand so in the same vein metaphors simileyou should absolutely use them becauseit helps people really step intosomething they understand already andthen you can describe on the technicalside sort of what things are actuallydifferent right what’s important topoint out that the simil metaphordoesn’t actually or would make you thinkright so like um basically like don’t beafraid of simplifying things as long asyou give people an understanding thatthat’s not the entire story but it canbe used to do what it can be used tohelp you understand how AI works it canbe used to you know explain that part ofittoo thanks for that yeah I I I I don’twant to give up on those metaphors so uhit’s it’s comforting to hear fromsomeone as expert as yourself that thatit’s okay it’s okay to use thoselanguage uh as long as we uh make surethat we also ground people in therealities um I I’m curious if you’veever encountered a situation where umwhere there was a miscommunication abouttechnology or a a place where you’venoticed a common place ofmisunderstanding and if you can share alittle bit about kind of how do younavigate you know when you’re kind ofdetective work digging into these toolsand understanding what’s actually goingon behind the scenes can you share withus a little bit about um youknow what what can go wrong but also howwe can recover from it yeah um it’s sofunny right so people in all sorts ofIndustries Opus skate what they reallymean for good and bad reasons all thetime we actually did a piece just acouple months ago about if you’rereading privacy policies now nowadayswhat what phrases mean secretlysomething else and how to decode that umso if you’re interested in privacypolicies you can totally check that outum but what I will say about um like thethe biggest maybemisnomer is to me and I say this all thetime which is artificial intelligenceright what AI stands for artificialintelligence is not artificiallyintelligent um instead to me it’sactually just really good branding so ifyou think about uh what the panel saidbefore me as well right it’s likeartificial intelligence sort of means alot of things um in a literal way but ifwe just think about it as the phraseitself it really communicates sentienceright and uh but there is actually nosentience yet but that implication sortof makes people sort of go down thisdoomsday sort of vibe Vibe of thinkingright and if you think about it there’sanother term we’ve thrown around both inthe panel and myself that was prettypopular right before artificialintelligence became the popular phrasewhich is machine learning right and ifyou think about the words machinelearning and then artificialintelligence they’re basically the samewords right they’re just each part ofthe phrases is a synonym for the otherpart artificial and machine learning andintelligent um andwhat I would say is that um when itcomes to AI I would kind of think of itas if it was a proper noun like cleanexinstead of tissue uh it’s not a brandobviously but it is now a standin rightfor something bigger and um we can useit that way because that’s just howpeople are literally using it but that’salso why the flip side of like you knowhow do we how do we get out of misnomersright is that when you are talking aboutAI overall I think you’re welcome to usethat and then once you get into thepracticalities don’t even use AI usewords that explain literally what’shappening right like um you know so andso Department uses a database withpeople’s information and matches it tothis in order to you know identifysomeone right just just describe what’sactually going on um and that that youknow comparison right there allows youto trigger in people’s minds like ohthey’re talking about AI That’s a bucketof information that have in my brain andthen here’s what’s literally happeningabout what I’m trying to communicate toyou yeah I feel like you’re uh I’m likeyou’re stealing my thunder here CeCe butI have been on a mission to just have AIstand for nothing my com’s director isnot a big fan of this but instead ofwriting artificial intelligence inparentheses Ai and then using AI throughthe rest of a piece I’m like can we justcall it AI like you were saying aboutlike you know calling cletex tissue ortissues Kleenex like like it what if itjust doesn’t stand for anything what ifit’s just a word and the word justhappens to be AI like that could be okayum right yeah I think it has taken on somuch more meaning um and and in that wayit is kind of it is I think even folkswho you know the the old school AIpeople you know the from like the 70sand the 80s uh folks who were working inthe space I think there’s some ation oflike oh people are using this to meananything these days but like but Iactually say as kind of liberating likepeople are using it to mean anything solike let’s use that as the bucket andthen get specific about the thinginstead of trying to like force peopleto only talk about AI in one particularway because it’s just cats out of thebag there I think yeah I think soSociety will take words wherever theywant to and if you fight against itpeople just won’t know what youmean yeah there’s definitely a duty umthe folks on the last panel shared youknow meeting people where they are is soimportant and so powerful and I thinknowhere is that more the case um than incommunicating to Broad audiences um inthe ways that journalists do so um Ireally appreciate that perspective I Iwonder if we might be a little bitself-serving here and ask you about someask you with your great editorialexpertise if you might uh talk a littlebit about something we wrote we sharedum with folks um these primers theseresources that we put together tointroduce people to the topic of AI uh Iwon’t say artificial to AI um and um andI’m curious if you might call out eithersomething that you appreciate in thoseprimers or something that you wouldchange or add or evolve about um the waythat we are talking about AI yeah umit’s a great question and it’s a greatresource and so if people are lookingfor avery quick and easy to understand primeryou should absolutely read it I willpull out one thing I really like aboutit because it matches the same type ofguidance that we give our journalistsall the time as well as you know anyonethat I’m talking to um this is similarto the metaphor or simile question butit’s a little different and it’s coveredin the primer two which is we neverpersonify algorithms or AI right soum this is something we’re verycognizant of at the markup before chatGPT or AI became a really popular phrasewe also don’t personify algorithms andthe reason behind that is aboutaccountability and agency right so ifyou say something like the AI did thisright then it’s suddenly like thisnebulous machine as opposed to a humanbeing use an AI to do something and thenyou’re like oh I I know how to navigatethat right and so um something that wetalk about a lot and I found I was justbrowsing through the primer and I waslike oh I found a very similar structureis that we always talk about who isusing AI to do what right and never theAI just does that because that’s not thefullstory um and so to give you an exampleright like college professors are usingsoftware that Flags whether a student’swork is AI generated right and then therest of our story is that except youknow the software almost always Flagsanything an ESL student writes as AIgenerated because it can’t tell thedifference um but you know in that worldyou’re suddenly like oh like Iunderstand how this fits into my lifepeople like my loved ones lives Societyand also I know who’s using it right andlike it comes in the form of software umit’s not like a mysterious AI machine isaccusing students of cheating rightthough I will say we are guiltysometimes of uh putting stuff like thatin our headlines to get people tounderstand that they should click andread the whole thing but then we werevery clear in the article exactly who isusing it and how but then also whatoptions do we have I think agency fromthe public is really important and um wecan talk more about that uh if you wantas well but just sort of this idea ofbeing clear about exactly what’shappening because it’s important andit’s important because it helps peopleunderstand not something that’s out oftheir control and there is a who that wecan hold accountable whether it’s aperson organization government entityEtc yeah yeah I I think I know you’vedone a lot of um work in the space andyou often call attention to um fantastickind of Labor organizers who are workingin the space of of interrogating and andum both leveraging and pushing back ontechnology um you know to the markupspoint challenging techn techology to doto to be better by us basically and um Ithink a common refrain we hear is likeyou know AI lays off whatever 5,000workers or AI is going to replace allthese people and it’s like it isn’t theAI that is doing that right like someoneat a company is making a decision to usea tool to decide who to lay off forinstance but right by by kind ofpersonifying and and placing so muchenergy um and attention on the tool wesometimes forget or um frankly hide thefact that there are people involved inthese decision-making processes and it’snot just the use of the tool whichcertainly is one area where people areinvolved but also there are so manypeople involved in building theseTechnologies and choosing what they’regoing to Priority prioritize andoptimize for so I think that thatguidance is is really resonant um for usand um I thank you for humoring me hereand toting our own horn um reallyappreciate your your expertise on thistopic yeah absolutely I I want to askyou and we’re goingon to Pivot to someaudience questions soon so flagging forfolks the Q&A button is down at thebottom of your screen please do shareyour questions um in the meantime I’dlike to ask you in the spirit ofcrowdsourcing like our Q&A um I’d liketo ask you because across your careeryou’ve used sort of crowdsourcingcitizen science from free the files tosome of the more recent citizen sciencestuff um tracking internet speeds um indifferent communities how can actuallythe public help journalists not justjournalists help the public but whatwhat can all of us do to helpjournalists do a better job umunderstanding reporting on um deliveringnews about these topics yeah um that’ssuch a great question there’s so muchthe public can do and and uh I’ll try toorganize these into just sort of twobuckets one is you can offer yourquestions right so if you don’tunderstand something I think you knowthis is something the journalismindustry is moving and changing towardswhich is really listening to readers andwhat their questions are and using thatto shape the questions that we ask ofthose in power or that we try to collectdata on because there isn’t any so ifyou have questions about stuff you knowyou don’t have to expect that say areporter will literally write an articlefor you but you shouldn’t hesitate toactually bring it up with whatever newsoutlet covers your area or the topicthat you’re interested in we get readerquestions all the time that we justanswer directly on social media um andit’s always common stuff that otherpeople would also want to know theanswer too right so on the questionfront I think uh just don’t hesitate toreach out I think it’s sort of sometimesjournalists can be sort of this um uhunknown and you don’t know how to interinteract with them if you’ve neverinteracted with a journalist before andin the same way that they call you upand ask you questions you can call themup and ask themquestions um but then more broadly rightI think that uh you can definitely offersort of your tips and tips is a wordthat’s used in journalism very broadlyto mean like if you have an idea for astory we should do like tell ajournalist because something’s happeningin your community and no one’s writingabout it right but what I mean here isthat I mentioned this earlier right in alot of our journalism at the markup wediscover the direct impact of atechnology and the community knows theimpact but they don’t know exactly howthe Technologies involved right and soif you feel like there’s somethingsuspicious or it feels kind of like at ascale that doesn’t make sense rightbasically if you have that thoughtthat’s like something’s wrong here and Iknow one part of the puzzle um that’soften enough information to help ajournalist start digging into it and tryto figure out what’s going on the otherthing and this is not even for howjournalists can cover the topics rightwhen we talk about how the public canget involved and um one of the thingsthat we have seen is that when Techmoves really fast right regulation hasto play catchup um and it just can’tmove as fast because on the one handthere are people who are dedicated toonly thinking about this topic and thenon the other hand they’re a Regulatorshave to think about all topics right soum something that we’ve been seeing isthat even more than legislators it’sactually unions that are creating someof the first real and practical rulesaround how AI can and cannot be used bycompanies and it’s from their bargainingand labor contracts right they’re sayingat our company this is what is reallyimportant to us and so there’s actuallya lot of um space right now for thepublic to participate in many differentways whether it’s about informationwhether it’s about um what they’reexcited about like should their kids beable to use it at school right you havea say in thatum but then also like how do I want mycompany to be able to use this becauseagain it’s not the AI that’s going totake our jobs right it’s the companydeciding that AI should take your joband um but it also is the companydeciding like let’s all use AI to makeour jobs easier right so there’s a lotif you think about it a lot that as amember of the public or as a member ofanorganization you can sort of interfacewith and advocate for things that you’reinterested in or ask for answers onquestions that you have and so I thinkyou know broadly speaking I generallyencourage people to engage with it andthen if you really want to helpjournalists um get at the heart of theseissues definitely feel free to reach outto them and you don’t have to have theanswer I think that’s the thing that Icouldn’t stress more you don’t have tosay I know exactly how all of this workseven if you know one piece of it rightthe journalist it’s their job to be ableto dig into therest so well said um and I I actuallyI’d love to I know this isn’tnecessarily the type of coverage thatthe markup does but I would love becauseyou mentioned it and actually one of ouraudience members asked a question um thewhile it is true that government tendsto move more slowly than some of thePrivate Industry it is also the casethat there’s been a whole lot going onin government action uh and talk anddecision- making and proposals aroundthese Technologies and one of ouraudience members asked what is youradvice for keeping up with this andunderstanding you know what what thestate of policy is you know in the US orglobally as it relates toAI yeah that’s a great question right itfeels like things are moving at 100miles per hour um and like yeah justreflecting what B just said right likeRegulators are paying a lot of attentionand trying to move as fast as possibleas well as people who are not ingovernment who are in nonprofits who areBanning together internationally even totry to like really set out someguidelines to help and so um this is oneof those things where in any topic ofnews it can easily feeloverwhelming and I think I have I havebasically like a couple of practicalrecommendations which is one don’t feellike you have to know the latest at alltimes that’s just just throw that outthe window that’s too much pressureabout any topic to put on yourself umand as journalists sort of we experiencethat pressure a lot because when youcover something you feel like you alwayshave to be the first one to know it evenbefore another outlet covers it soyou’re talking to the right people Etcjust leave that leave that to the sideum what I would do is I would thinkabout what kind of news have youconsumed that you sort of like thoughtwas really helpful right so like thearticle you read yesterday or thepodcast you listened to and what I woulddo is try to TR to find like one gobeyond one but go to like one two threesources let’s say And subscribe to theirnewsletter or something like that andthen just like if if your goal is to beup to date to just read what isinteresting to you that is sent to youin those newsletters right sonewsletters often create like a menu foryou to choose from where there’s tons oflinks and you can just read a sentenceabout each one and they will do all thecuration work for you about who thatnewsletter curator whose work they thinkyou could read today that’s new rightand I find that that is sometimes themost manageable way to really stayinformed without feeling like oh Ididn’t read the thing that just came outbecause I guarantee you something big isgoing to be in every newsletter likethat that’s no question um but there arecertain elements like if you really likea publication just use their newsletteras a way to help you manage informationoverload Absol absolutely got to plugthe markup as a fantastic source ofinformation um to stay up with and II’ll be Shameless here and plug ournewsletter too check check out the asdigal newsletter um thank you so muchfor that uh really really detailed umand specific answer which is soappreciated um as we as we wrap up hereI want to bring one more audiencequestion into the chat uh before we letyou go um which is you know folks shareda ton of re resources here there weretons of links shared earlier oh mygoodness um we will be sharing thoselinks in the followup from theseconversations um but you know inaddition to all the resources andorganizations and Tik toks to watch umwould love to hear from you what do youwish was getting more attention what’ssomething that you’re really fascinatedby that you’re digging into or a topicthat um you think you know or a spacewhere there’s not technology being usedbut you think that there are to bewhat’s something that um you think thatyou could shine a little more light onfor us o so many juicy topics that Icould bring up hereum okay I’ll just pick I’ll pick one butI think I think one of the um mostbefuddling topics to cover and has beenfor the long longest time is probablyprivacy right so when it comes to yourprivacy online there’s a ton ofresources tools Etc right how to makeyour email comes texting Etc everythingprivate but the hard thing aboutcovering privacy is that it feels like abig brother moment where you’re like youknow this could happen it might behappening but it’s really hard toconnect all the dots that’s like as aresult of not having privacy in acertain area this is now the consequenceright that last mile is extremely hardto figure out and so generally you havepeople who are concerned about privacyor not um and when it comes totechnology right it often has to do withlike am I willing to use this technologyeven though I know that there’s somelikeunknown situation I’ll be in because I’mgiving up my personal information rightand like it just feels like this fogright like people don’t know how toreally quantify it or like tangibly holdit in their hands and what’s interestingabout Ai and privacy right is that thereare are some really clear tangiblethings such as you know if we’re if inset of AI we’re talking about as largelanguage models and chat gbt it is usingall the information on the internet uhincluding your information right likeartists have really um tangible ways ofthinking about it which is if it’sconsuming their art and trying to learnhow to do art from it is it a derivativeall that stuff and when it comes to dataprivacy there are ways for you to peoplehave written tools for you to be able tobe like is my website being used totrain Chad gbt right um and I thinksomething that is not getting enoughattention though I’m we’re we’rewrapping up our coverage of it a littlebit as well is that people have alsobeen trying to figure out a way suchthat like all things in privacy couldthe development of AI and how much ituses your data be an opt-in situationand do you have a way to opt out rightand so there actually there’s actuallycompanies and policies in developmentout there that’s about like if you don’twant your information out there how canyou hide it essentially from people thatyou don’t want to just take it for freewithout your consent um and that I thinkis something where I feel like it’s oneof those um Wicked problems right wherewe all know that if we could just beomniscient and see everything about theissue there would be so many revelatorythings but because it’s hidden the bestwe can do at the markup as journalistsas the public is trying to make astransparent as possible what we do knowand try to like bring more stuff tolight because there’s so much moneymoving around for your data um and uhsomething is always difficult when it’sin this nebulous foggy space as opposedto a tangiblespace yeah for sure I think this is I Ican’t wait CeCe to have another chatwith you because this is a topic I’mvery passionate about um topic I knowour team is very passionate about um andreally grateful to everyone whosubmitted in the audience Q&A um to tobring us these great questions for ccway um CeCe what a joy so much knowledgeso much wisdom um thank you so much forjoining us as we transition here we’regoing to um have a special announcementum so I’m very excited for what’s nextbut I just want to invite everyone togive a big hand to Cece and to all ofour panelum for making this conversation possiblethank you thanks for havingus
thanks everyone uh for joining in wehave as I mentioned a specialannouncement and I’m going to turn itover to my fantastic colleague um Tomlowski who’s going to talk us throughwhat we are announcingtoday all right well thank you so much Band thank you to CC and all of ouramazing panelists today for being heresuch a such an exciting event um ifthere’s been one central theme today ofmany it is a lot of uncertainty aboutwhat the future will look like and howAI technology will develop just tohighlight that a little bit StephenHawking said that AI could spell the endof the human race Mark Andre thinks AIcan make everything we care about betterHillary Clinton has said we’re racinghead first and we’re unprepared Yan laonof meta has said we should not see AI asa threat we should see it as beneficialvice president Harris thinks that AI hasthe potential for profound good andprofound harm and President Barack Obamahas said it’s clear by now that AI willaffect us all and I share this to reallyemphasize that people disagree there’snot one unified Narrative of what thefuture is going to be like it peoplehave opinions spanning from the end ofthe human race to everything is fine andeverything we’re going to live in aUtopia to profound impacts on Democracyon the labor force to profound positivesthat could come to Consumers to all ofus and so what does this mean it means acouple things first it means that weneed to be thoughtful and it also meansthat we need to make sure that a lot ofpeople have a voice that everyone canhave a voice in this conversationbecause we don’t know what the futurewill look like and so we need to listencarefully to one another and so this isone of the central drivers of Aspendigitals work and so as we’ve mentioneda couple times um one of the centralthings we’ve done in this vein is ouremerging Tech primers um which you canaccess at this URL and in the chat um westarted the process of creating theseprimers over a year ago they startedwith interviews with journalists whichtransitioned into a series of salonswith experts on AI and eventually wepublished three AI primers reallyinitially intended to help journalistsas they figure out how do we report onAI how do we talk about AI they wereboth designed for Tech journalists andalso for non-tech journalists right forthe folks as we heard CC mention whocover other beats but who arerecognizing well AI is affectingHealthcare or AI is affecting Finance orpolitics and really need to be catchingup on these newtechnologies so I’m going to quickly gothrough those three primers firstthere’s finding experts in AI this hasseveral components one is informationabout common roles in AI one is adirectory of experts so how do you getin touch with the right people who doyou need to talk to to learn more aboutAi and as well as we have notableconferences in AI right so so where arepeople Gathering where are theseconversations ation happening second oursecond primer of three is the intro togenerative AI this highlights what isgenerative AI how does it actually workwhat does it mean and also goes intoseveral key issues so those include thefuture of work intellectual propertyinformation ecosystems the effect ofgenerative AI on our ability to talk toone another and to have a shared senseof Truth and severalothers and finally our third primer AI101 this goes into the very ground levelwhat is AI what are we even talkingabout what does all this mean people saythe algorithm what does that mean whenpeople say AI what does that mean and wealso discuss how to talk better about AIso here is our view of what the bestsentences on AI look like we think theygenerally take the form of people whouse specific AI tools or capabilities todo tasks so just to highlight each pieceof that first people right so we’re notnot just talking about AI as an agentwe’re talking about the people whodesigned the system who trained thesystem who are deploying the system andmaking those decisions about how it getsused second specific tools orcapabilities we think it’s important toget away from talking about AI as onething and it’s far more helpful to talkabout specific tasks specificcapabilities that AI has and finally thetask portion we want to get away frommarketing speak fromyou know projections about what thingscould look like we want to talkspecifically about what AI is doingnow so I’m going to run through anexample of what a good excerpt about AIlooks like this is a bad excerpt aboutAI I wrote this excerpt about AI um andwe’re gonna go through it and improve itso right now it says the AI has learnedto predict crimes how can this improvewell first has learned this ispersonification right we use thissometimes to help explain to to folkswho are new what’s going on but this isnot actually what’s happening within thesystem right so we can get more specificso this is a little bit better the AImatches crime data with informationabout land use to predict crimes that’sa little bit better but we can keepimproving right now we’re still talkingabout the AI as if it’s one thing we canget a little bit more clear there riskterrain modeling matches crime data withinformation about land use to predictcrimes that’s even better but we cankeep improving predict crimes this ismarketing speak right is that actuallywhat’s happening what does the systemliterally doing we can make this excerptbetter and here we go risk terrainmodeling matches crime data withinformation about land use to identifytrends that could be triggering crimesthis is even more clear but we can do alittle better who is doing this we’restill centering the AI the risk terrainmodeling as the agent here so we canimprove this further by talking aboutwho is actually creating this who isdeploying it so here’s the final versionof this sentence developed by ruerUniversity researchers risk terrainmodeling matches crime data withinformation about land use to identifytrends that could be triggering crimesand CC if you’re still on this sentencemight look familiar because this is froman article in the markup we think thisis an an excellent example of how totalk better aboutAI again what we think this sentencedoes well is it talks about people whouse specific AI tools or capabilities todo tasks so just to highlight that herewe have the people behind the tooldeveloped by Rucker Universityresearchers we have the specific toolthat’s being used risk terrain modelingand we have the specific capabilityavoiding marketing speak talking aboutwhat it’s actually doing it’s matchingcrime data with information about landuse to identify trends that could betriggering crimes this is an excellentexample about how to talk about AI wewant to highlight examples like this andso as part of that effort today we’revery excited to launch our reporting onAI Hall of Fame if you have written anexcerpt about AI or you’ve read anexcerpt about AI in an article we areasking you to submit at this URL wewould love to highlight these exampleswe want to point people towards the bestexamples of reporting on AI so just toquickly run through what we’re lookingfor first we’re looking for Clearlanguage we want these examples ofexcerpts to be accessible to as manypeople as possible we don’t want it touse technical jargon second we’relooking to highlight human actors so sowe again think it’s important that theAI is not the agent in these EXs we’relooking for the humans that aredesigning the systems that are trainingthe systems and that are deploying thesystems third we want to describecurrent capabilities so we’re trying toavoid marketing speak in these excerptswe’re trying to describe specificallywhat current systems aredoing and finally and this is Central wewant to avoid personification right sowe think it’s really important to talkabout what the AI is literally doing notusing uh you know figurative languagethere is a time in place for that wordslike learn or understand sometimes canbe helpful when folks are are new but wethink it’s important that these excerptsmake clear what is literally going onwithin the system so again we’re askingall of you if you’ve written somethingif you’ve read something whatever submitto our AI Hall of Fame you can do so atthis link below and hopefully it’s inthe chat as well our deadline forsubmission is the end of January so youhave until January 31st 2024 again thiscan be something you’ve written can besomething you’ve read as long as it’s anexcellent example of AI writing we wantto highlightit thank you very much I will pass itback toB thank thank you so much Tom and thankyou everyone for joining us here todaywe are so excited to be launching theHall of Fame and I’ll just say again weencourage a wide range of submissionshere are you uh involved with a studentjournalism Outlet are you involved witha school newspaper we would love tohighlight great technology journalismthere if you are someone who is umsharing around a sharing about a Englishlanguage um publication from around theworld we would love to include as manydifferent examp examples of greatdescriptions of AI that we can find soum please do this and the reason we’redoing this is because what we heard whenwe did those initial interviews withjournalists back in the DayDay when wefirst started this project what we heardfrom folks is I need good examples tomodel after I need templates I canfollow I need examples that showcasewhat good reporting and goodcommunication looks like and as weexpanded this work out from journaliststo reach broader audiences in governmentin nonprofits and education what wefound is the same is true everyone needsgood examples so by contributing yourgood examples um we can make uh make theconversation about AI a little bitbetter so with that I want to bring usto a close here thank you so much to ourpanelists Victoria Alex Clarence CCE forour fireside chat absolute Joy um it’sbeen so wonderful having you all herewith us thank you all everyone joiningin from home um who participated in ouraudience Q&A really appreciate all ofthe thoughtful questions that you allsubmitted looking forward to the nextone and I hope that you will join uswelcome you to follow us on LinkedIn Ibelieve we have a Twitter account um wehave uh uh our team and our paneliststagged in our LinkedIn posts um andabsolutely please do sign up for ournewsletter so you can stay tuned for allof the upcoming events a big thank youto our funders at seagull familyendowment who’ve been incrediblePartners in this work we couldn’t do itwithout you thanks so much everyone hopeyou have a great rest of yourday
Speakers
Victoria Houed Director of AI Policy and Strategy, U.S. Department of Commerce
Read about Victoria
Victoria is an engineer turned policy entrepreneur who has dedicated her career to improving the tech policy ecosystem and the modernization of government. In 2023, she joined the U.S. Department of Commerce as Director of AI Policy and Strategy.
Victoria was previously a Manager at Schmidt Futures where she funded key science and technology policy initiatives and educated relevant stakeholders on how to best advocate for their policy ideas in the legislative and executive branch. Previously, Victoria worked for House Speaker Nancy Pelosi as a TechCongress Congressional Innovation Fellow supporting the Speaker on all things tech, including internet accessibility, data privacy, antitrust, disinformation, Section 230, and autonomous vehicles. Before her time in Congress, Victoria worked as a Software Engineer and Product Manager at Cards Against Humanity, founded a non-profit for Black Women in Technology called BlackByte, helped lead the tech team of a mayoral campaign, and enjoyed working on different engineering projects for artists around her hometown of Chicago.
Victoria holds her B.S. in Industrial Engineering from Purdue University, where she focused on nanotechnology and systems engineering.
Alex Kotran Chief Executive Officer, AI Education Project
Read about Alex
Alex oversees strategy, partnerships, fundraising, and external relations for the AI Education Project. Prior to founding the AI Education Project, Alex built the AI Ethics and Corporate Social Responsibility function for H5, Inc., a pioneering AI company in the legal services sector. At H5, Alex led strategic partnerships with organizations including the United Nations, NYU School of Law, the OECD, and IEEE to develop judicial education programs and high profile convenings. Prior to his time in the social impact and non-profit space, Alex managed brand and policy communications for companies including Oracle, Airbnb, Nissan, HP, Adobe, and SAP. He was a lead Field Organizer in Columbus, OH for the 2012 Obama Campaign and served as a Presidential Appointee under HHS Secretary Sylvia Burwell, where he managed communications and community outreach for ACA Open Enrollment, two SCOTUS rulings, and the 2015 Ebola outbreak.
Clarence Wardell III Senior Program Officer, Bill & Melinda Gates Foundation
Read about Clarence
Dr. Clarence Wardell III is a changemaker committed to harnessing the power of technology, data, and evidence to create positive social impact and achieve more equitable societal outcomes. He is a Senior Program Officer on the Economic Mobility and Opportunity team at the Bill & Melinda Gates Foundation, where he leads the team’s efforts to foster innovation to improve economic outcomes for individuals in the U.S. with low incomes. Before joining the foundation, Clarence served in several senior roles in the Biden-Harris Administration and on the transition team.
Clarence has also previously served in roles at Results for America, CNA Corporation, and in the Obama Administration. Throughout his career, Clarence has been actively involved in various organizations focused on economic opportunity and criminal justice reform as a board member and advisor. He holds a B.S.E. in Computer Engineering from the University of Michigan and a Ph.D. in Industrial and Systems Engineering from the Georgia Institute of Technology.
Sisi Wei Editor-in-Chief, The Markup
Read about Sisi
Sisi Wei is the editor-in-chief at The Markup, a nonprofit investigative newsroom that challenges technology to serve the public good.
Before joining The Markup, she was co-executive director of OpenNews, where she envisioned and executed transformative initiatives for journalism. As part of her work, Sisi founded the DEI Coalition, a journalism community dedicated to sharing knowledge and taking concrete action in service of a more anti-racist, equitable, and just journalism industry.
She was assistant managing editor at ProPublica from 2018 to 2020, where she oversaw three editorial teams focused on news apps, interactive storytelling, and visual investigations. She also managed large, interdisciplinary investigations across the newsroom, one of which won the Pulitzer Prize for National Reporting in 2020. Sisi worked at ProPublica for seven years, investigating abuses of power and betrayals of the public trust across a range of topics, including health care, higher education, government, and immigration.
Sisi also serves on the board of News Revenue Hub and The Open Notebook.
Moderators
B Cavello Director of Emerging Technologies, Aspen Digital
Read about B
B is a technology and facilitation expert who is passionate about creating social change by empowering everyone to participate in technological and social governance. B is the Director of Emerging Technologies at the Aspen Institute.
Previously, they advised Senator Ron Wyden on issues of privacy, internet governance, and algorithmic accountability through the Congressional Innovation Fellowship. Prior to Congress, B was a Research Program Lead at the Partnership on AI, a non-profit multi-stakeholder initiative to address the most important and difficult decisions on the future of artificial intelligence. B led research in areas such as fairness, transparency, AI’s impact on labor.
B has also served as Senior Engagement Lead for IBMWatson, where they led strategic conversations with world leaders about AI and patented a spooky authentication technique. They were also recognized as an IBM LGBT+ ‘Out Role Model’ for championing diversity and inclusion in and beyond IBM. B earlier served as Director of both Product Development and Community at Exploding Kittens, a record-breaking crowdfunded card game startup. B also co-led Phenomenon Media 501c3, and arts and education nonprofit, developing an educational tool for teaching code without the need for computers.
B received a Bachelor of Science in Economics from the University of Texas at Dallas and was selected as an MIT-Harvard Assembly Fellow for the 2019 Ethics and Governance in Artificial Intelligence Initiative cohort.
Eleanor Tursman Emerging Technologies Researcher, Aspen Digital
Read about Eleanor
Eleanor Tursman is an Emerging Technologies Researcher at The Aspen Institute, where they work at the intersection of novel technologies, public education, and public policy. They are also joining the Siegel Family Endowment as a Research Fellow for 2022-23.
Eleanor is a computer scientist by training invested in bridging the gap between academic research and policy to promote social good. They believe in advancing science-based governance of artificial intelligence and other emerging technologies, tools which—despite their potential benefits—can also be used to disenfranchise and discriminate against minority groups, threaten democracy, and propagate disinformation.
Prior to joining The Aspen Institute, Eleanor served as a TechCongress fellow in the office of Representative Trahan (D-MA-03), providing technical and scientific support for topics including health technology, educational technology, artificial intelligence, algorithmic bias, and data privacy. They also worked as a computer vision researcher at Brown University, with a specialty in 3D imagery and the human face. Their research centered on finding new ways to approach deepfake detection, with a focus on long-term robustness as fake video becomes harder to visually identify.
Eleanor has an M.S. in computer science from Brown University and a B.A. in physics with honors from Grinnell College.
Tom Latkowski Google Public Policy Fellow
Read about Tom
Tom Latkowski is an MPP candidate at Georgetown’s McCourt School of Public Policy who currently works as a Google U.S. Public Policy Fellow at the Aspen Institute. Tom has previously interned at the White House Domestic Policy Council and the Office of Senator Dianne Feinstein. Tom has previously worked on campaign finance reform, including writing a book on democracy vouchers, and co-founding an organization to advocate for campaign finance reform in Los Angeles. Tom attended UCLA for his bachelor’s degree, where he studied Political Science and Applied Mathematics.
{"includes":[{"object":"taxonomy","value":"131"}],"excludes":[{"object":"page","value":"203292"},{"object":"type","value":"callout"},{"object":"type","value":"form"},{"object":"type","value":"page"},{"object":"type","value":"article"},{"object":"type","value":"company"},{"object":"type","value":"person"},{"object":"type","value":"press"},{"object":"type","value":"report"},{"object":"type","value":"workstream"}],"order":[],"meta":"","rules":[],"property":"","details":["title"],"title":"Browse More Events","description":"","columns":2,"total":4,"filters":[],"filtering":[],"abilities":[],"action":"swipe","buttons":[],"pagination":[],"search":"","className":"random","sorts":[]}
Facebook
Twitter
LinkedIn
Email
