good morning or good afternoon thank youso much for being heremy name is betsy cooper and i am thedirector of the aspen tech policy hub weare an incubator training technologiston how to get more involved in policythrough policy training programs andchallenge prizes such as today’s eventson behalf of my colleagues at aspendigital and the commission oninformation disorder we are delighted tokick off this live information disorderprize competition by the end of today’ssession we will have awarded 75 thousanddollars to one of our amazingcontestantsthe commission on information disorderwas created to identify the mostcritical sources of miss anddisinformation and deliver a set ofrecommendations to help government theprivate sector and civil society respondto this challengefollowing months of internal discussionand consultation with experts thecommission comprising a diverse grouprepresenting academia governmentphilanthropy and non-profits released areport detailing 15 actionablerecommendations to combat informationdisorder these recommendations fall intothree categories increasing transparencyand understanding building trust andreducing harmsthe aspen tech policy hub with supportfrom craig newmark philanthropies andex-ante an initiative of schmidt futureslaunched this information disorderchallenge grant to fund unique andinnovative projects that make meaningfulprogress towards ending informationdisorder indirect connection to one ormore of the commission’s 15recommendationsproject teams were invited to identify aparticular recommendation from thecommission’s report and propose a newuntested solution that wouldspecifically help accomplish this goalwe selected four semi-finalists that weare excited to present to you here todaywe awarded them each five thousanddollars to develop prototypes of theirdeliverables over an eight week periodand today is the day each team will bepresenting their accomplishments andexplaining how their contribution is themost unique and innovative projectmaking meaningful progress towards andending information disorderthe projects will be judged live afterwhich one team will be awarded a 75 000grand prize towards executing itsproposed ideaour judges for this event are three ofthe original commissioners from thecommission on information disorder kriskrebs amanda zamora and debroythe run of show is as follows each teamwill have five minutes to demonstratetheir proposed solution the judges willthen have five minutes to ask questionsof the team if you as the audience havea question that you think the judgesshould ask feel free to submit it in theq a box at the bottom of your screenafter each team finishes presenting thejudges will depart to a separatedeliberation room while vivian schillerand ryan merkley of aspen digital willprovide an update on the commission’swork after deliberations the judges willreturn and announce the winnerso without further ado we’d like tointroduce our first team alteria inc anexperienced design company based in losangeles alteria creates live and digitalexperiences that draw on history and artto explore the theme of agency todaythey’re demoing agents of influence amedia liter a media literacy video gamethat teaches middle and high schoolstudents to recognize misinformationthink critically and make moreresponsible decisionsrepresenting the team today are anahitadalmia and jasper mcavoy over to youteamthank you so much for the kindintroduction betsy whenever you’re readyto ask for himand just to clarify we have 10 minutesto present not five right just want tomake sureyes you have 10 minutes to present fiveminutes for questions apologies okaythank yougood morning everybody and thank you somuch for having us over here today i’mgoing to be sharing with you agents ofinfluence a spy themed media literacyvideo game that teaches the nextgeneration to combat misinformationthe term actually originated in the coldwar era where agents of influence wasused to refer to spies that spread missand disinformation in foreign countrieswe wanted to reclaim the term to showpeople that their agency their abilityto navigate the world around them andtheir influence their ability to make adifference are positive traits for us itstarted during the pandemic when werealized that we were at a historicalpivotal pointevery decision we made seem to mattermore than ever whether to wear a mask ornot to bear a mask how you speak aboutthe black lives matter movement who youvote for we were all contributingtowards making a new normalthat’s when the problem of miss andmisinformation came to our attentionas we realized that it was disempoweringpeopleby misleading them into taking decisionsthat led to un unwanted outcomesit is for that reason that we createdagents of influence and let’s look atthat[Music]welcome to agents of influence cyberdanger an educational spymedia literacy video game created toempower the next generation of studentsto fight misinformation on the internetvirginia hall high is under attack anefarious organization named harbingeris manipulating the student body withlies and deception luckily you and yourteam the agents of influence are thereto stop themstudents will learn to question thetrustworthiness of informationinvestigate the trustworthiness ofinformation and apply this knowledge tomake better informed decisionsit was during this design process thatwe realized that misinformation was apressing global problemand our vision for this game is thatit’s used in classrooms across the worldempowering the next generation ofdigital citizensour goal is that by playing this gamestudents learn to realize the impactinformation hason their worldview and decision makingand have the ability to find theaccurate information so that they canmake decisions that truly leads to theworld they want to see we also intendfor this game to be a doorway to a hostof civic education and civic engagementin many ways the same way the aspenreport was a doorway for uswe realized how complex the informationecosystem we were living in wasthat we were just a small cog in thismachine a sub-category of a sub-categorybut we also recognize that by empoweringpeopleparticularly the next generation we aregoing to have a positive ripple effectin all of the other areas of aspen’srecommendationsthat was one of the factors that weconsidered while picking our targetaudience which is 13 to 15 year oldsthis audience is also at a criticalstage of development where they’rebuilding lifelong habits andthey’re always onlinethis has inherently changed the way theyinteract with the world around themthis generation expectsconstant engagement and immediatefeedback making them particularlyreceptive to gaming as an educationalapproach research also shows that gamingis a very powerful educational tool asit is highly transferableprovides immediate and specific feedbackallowing people to change their habitsin real time and is engaging which meansthat itresonates emotionally is highlymemorable and makes students actuallywant to learnthis is all that so now that we’veestablished that let’s have a look atthe story world of our gamelife moves fast at virginia hall highschoolwhen i first arrived that was just anormal freshman now i’m a spya few weeks ago a shadowy figure knownonly as harbinger started causingtrouble at schoolharbinger used a device called thelullaby to transform some of thestudents into a kind of mind-controlledbeing that can influence the thoughts ofotherswe call themsleepwalkersthese altered students are dangerousassets in harbinger’s plots to spreadmisinformationluckily i’ve joined with a group ofstudent secret agents to stop them wecall ourselves the agents of influencein this adventure students are going tobe encountering three main challengeswhich take the form of our core gamesand are where they’ll find majority oftheir media literacy skills the firstgame is the conversation game one of themost negative impacts of misinformationis the polarization we are facing todayand that people have forgotten how tohave healthy conversations with peoplethey disagree with this game aims tosolve that and also works towardsanother aspen recommendation of healthydigital discoursein this game students are tasked withinvestigating a suspect and by and theydo this by picking dialogue choicespositive dialogue choices such asleading with empathy will get thisperson to open up but negative dialoguechoices such as being critical blamingor being defensive is going to get themto stop talking to youour next game is the analysis gamestudents are overwhelmed we all areoverwhelmed with information today andwe’ve stopped looking critically at theinformation we’re consuming the analysisgame helps us overcome that in this gamestudents realize that their secretoperatives hiding in underground tunnelsbeneath their school to reach them theyneed to overcome a security system bycorrectly answering true false questionsabout an article they read and reading astrat and playing a strategy game on theleft side through this students learn totell the credibilitybias and purpose of the content they’reconsuming but also recognize thedifference between the main idea andnuance in this content because today thesame content can be propaganda marketingand entertainmentall at the same timeour final game is the research gameand this is based on social media andthe internet which is where people findthe majority of misinformation in theirdaily life so it can be highlytransferable students are tasked withfinding culprits for spreading miss anddisinformation on their school socialmedia by researching posts further andtagging them as accurate or misleadingin this process they tell the differencebetween fact and opinion and crossverify information across sources tofind trends discrepancies and makejudgments on what information theyshould be trustingif you successfully complete all ofthese games you will have beat harbingerand saved your school frommisinformationhowever even though the game ends thenyour engagement with the topic does notwe intend to give studentsmisinformation toolkits which includestools such as fact checkersbias test and partner curriculum werecognize this is not a problem we cansolve alone and to and want to work witheveryone else in the ecosystem to getmaximum resources out there as easily aspossible we also want to give studentsthe opportunity to pitch real worldsolutions to real world problems throughthings such as design competitions andwant students to feel empowered to makedecisions on accurate information thatwill lead to the world that they want toseewe expect these students to becomeadvocates for the truth sharing whatthey learn with their friends and familymaking a positive impact on theircommunity and having a ripple effect onall of aspen’s recommendation butcreating the next generation ofresponsible digital citizensour goal over here is to have over 100000 students playing by 2024 and you’realready well on our waywe have three school districts incalifornia ohio and pennsylvaniacommitted to piloting this game bringingus to about 10 000 students with furtherdistribution partnerships withorganizations such as rotary and teacherassociations we also want to ensure thatthis game is effective so we haveresearch partnerships with wested andstanford to get feedback and continueiterating on the gamethank you so much for taking the time tolisten to us we look forward to workingwith everybody in this room to solvewhat we recognize as one of the mostpressing problems of our eraback to you betsy happy today questionsfantastic so now i’m going to ask thejudges to come up and judges you havefive minutes to ask the team questionswho would like to go firsti have a quick questioni was first of all this is fantasticthank you so much um for taking the timeto put this togethermy question is just to understand alittle bit more about what your outreachstrategy is to build participation inschools you’ve got these three pilotslike what are the measures of successthat you have with these pilots and howdo youexpect toto expand i guess in other districts orleverage other community partners to toreach your goalwe have once we have some slides on thiscan i uh start sharing screen again ordo you want me to take it off screenshare um whateverokay cool i’ll put it back on screenso we have recruited over 500 studentsbeyond these pilot programs for artistssome of our metrics include how muchstudents care about the topic ofmisinformation whether they’re going tolike research before they shareinformation again how confident theyfeel about disowning misinformation anda number more if you want i’m happy tosend you our entire research plan afterthis which wested and stanford haveworked with us to designum we have a number of distributionpartnerships which include medialiteracy curriculum rotary clubs we’vedoneover 40 presentations already and manyschool superintendents are rotarians sowe get to chat with them directly oninterest in this game um teacherassociations including nationalassociation of media literacy educationenglish teachers the writing project andmoreand once we have proof of conceptthrough this pilot game and effectiveresearch many people have said come backto us to share this game so we can takeit you know mainstream and we’ve alsodesigned this game flexibly so it can befit in a major like a variety ofcurriculum english class media literacyclass after school curriculum and morefor it to be as widely usable aspossibleother questions from the judgesyeah um so i have a uh one kid at leastthat’s right below your kind of targetaudience i have a 12 year old umand seeing how they play games uh ithink gamification is a great way toengage i guess my question isfrom a from a gameplay perspective isthis ayou sit down and you play it for 30minutes you play it for five hours kindof what’s the user experience andunexpected uh is it one time is itmultiple times you keep coming backso something we learned was reallyimportant is flexibilitywe hope for students to start playingthis game in the classroom and our twohour module first two hour module willbe broken down into 30 minute segmentsthey can play them independently or theycan play all four bringing them to theultimate narrative conclusion the goalis that they get invested in the storyso even if they don’t finish at schoolthey’ll finish at home we intend to makeseveral more modules putting us at over10 plus hours of gameplayso that students can engage with thistopic holistically and the lessonsactually are deeply seated when they’relearning rather than something in ourinformation era that they check in onforget about and move ongreat thank youi’m curious umwhen you talk aboutyou knowlooking for accurate sources just howare youthinking offor let’s sayif there’s some skepticism aboutthe the game andsome parents saying i don’t wantthis as a basis for my kids formingtheir you know kind of beinginfluenced in ways i justi question how will youdefine accuracythat sort of underpins thestructure of the gamethat’s a very good question particularlybecause that’s a problem we are alreadyencountering the way we’ve been tacklingit is that we’ve been emphasizing thisgame is skills oriented we are nottouching upon any real world problems inthe game itself you’re going to betalking about school elections andbetting school sources in a fictionalworld so as long as we are teaching themthe skills correctly how do you researchcheck multiple sources have differentperspectives can you recognizecredibility can you recognize bias theycan’t say that this is inaccurate and weunderstand there is a lot of nervousnessaround the word misinformationparticularly because of the tension andmisinformation free speech and the factthat many times things exist on a scaleof truthfulness we want to encouragethose conversations sometimes you can’talways find what’s most accurate even ifyou know the knowledge exists so theinformation exists out there that’ssomething to be conscious about in yourdecision making and we want to give youtools resources skills and mostimportantly a concern for theinformation landscape you’re living inso that you’re best equipped to makeresponsible decisions with all theinformation you have on handfantastic and that’s time thank you somuch to the alteria team for presentingwe really appreciate itnext we’d like to turn to our secondteam ratitude ratatube seeks to helpusers discover search through andanalyze hard to find youtube contenttoday they’re showcasing a tool toautomatically detect and trace harmfulnarratives spread through youtube videospresenting today are cameron ballard anderic zimmerman over to you cameron andericall right um hi everybody thanks foryour time my name is cameron and i’m aphd candidate at nyu tandon’s cybersecurity for a democracy center i’m herewith eric who’s a technology fellow athuman rights innovation lab and we wantto talk about a tool we’ve been buildingto discover and trace the spread ofharmful narratives across youtube and sothis whole thing started when eric and iwanted to look into content moderationon youtube at the start of the covet 19pandemic we knew that youtube was a bigsource of disinformation includingconspiracy thinking extremist contentand foreign influence operations but wethought it got a lot a lot lessattention than sites like facebook andtwitter when it comes to disinformationum and sure enough when we startedlooking into content moderation werealized that there aren’t very manywidespread research tools for this kindof analysisand we think this is due to two mainproblems the first being that there’s somuch content on youtube and audiovisualmaterial is difficult to research so atwo hour long video has a lot more kindof fluff but also a lot more contentthan the 140 characters of a site liketwitter as well the recommendationalgorithm and opaque search engine makeit difficult to find content that you’relooking for since so much of contentdiscovery on youtube is driven byalgorithms it’s very hard to findcontent outside of a community thatyou’re not already a part of and so wedesigned the first prototype ofgratitude which keepswhich indexes the transcripts andmetadata of a bunch of differentchannels a number of which had beenidentified as radical or potentiallyharmful by previous research uh now butwe recognize that that’s not enough tosolve the kind of complex problem ofdisinformation on the site and so we’vebeen building a tool that mergesqualitative and quantitative researchmethodologies to trace narrativeswe think this goes a long way towardsaddressing a lot of the recommendationsin the commission on informationdisorders final report but especially inareas of increasing transparency andharm reduction and so by having thiskind of data public and easilysearchable uh we can not only increasepublic interest research of uhdisinformation on on youtube but we canalso help verify any sort of contentmoderation that youtube may or may notcarry out as well by figuring out howthese narratives spread and how muchthey’re driven by major content creatorsversus other methods ofdisinformation spread um we can add moreaccountability to super spreaders andmaybe counter these narratives beforethey cause any real harmin fact our tools already being used toaddress some of these problems um and wesee it being useful in a wide variety ofuse cases including but not limited tounderstanding foreign influenceoperations uh increasing electionintegrity um better platform auditing ascalled for the new eu digital serviceact as well as combating extremism andincreasing ad transparency uh in factit’s already been used by researchers atthe shorenstein center looking at uhvoter fraud claims during the 2020election uh asian americans advancingjustice group used it to find instancesof anti-asian hate and calls forviolence which they presented in frontof a congressional committeeas well the times london used it to findanti-vaccination videos um that uh thatare still being monetized throughadvertising on youtube uh recently we’vebeen working with a group calledsparvarius to fight white extremism onthe site um and as part of that we’vebeen building a prototype for aspenwhich erica is going to walk you throughnowyeah so to give a specific example onhow raditube tries to solve for case offertilization i’d like to show thisresearch collaboration we did withbavarias and for this we’ve beenspecifically looking into media to riseand the affiliated militia group pagesfronts and mediatorize is a propagandaoutlet that’s promoting the white powermovements to a series of documentariesand video reports and one of theirproductions is exposing the truth inwaukesha documentary which exploits andframes the waukesha christmas breakattack as an anti-white terror attackall their official youtube channels havebeen deleted from the platform butyoutube is still being used to reach anew audience which they cannot reach onalt tech platforms in one telegrammessage the groups ask the users toupload backups of the filmon the disguised titlesand we can use reddit to discover thosevideos and see how they spread to theplatform and for this we need a coupleseed videos and channels which you foundon the telegram channel and which wethen submit to the crawler and just tomake clear we made this this front endduring the the competition butparts of the back end we already hadand under the hood this tools works bylooking at the comments and publicsubscriptionsin the past youtube has sometimes beenseen as more of a video platform thanthe social media websites but there’sactually a lot of interaction going onbetween users as well as users andcreators but also between creators thatinteract with each other to videocollaborations or video responsesand a recent survey also shows thatyoutube is actually the most popularsocial media platform in the us and isused by 81 percent of the populationsoto go back to to our results once thecrawler is done we can find thosechannels that are being watched by thesame audienceand through both the graph as well asthe list we are able to find channelsthat are in close proximity to eachother we use the three-dimensional graphhere which does look really pretty butwe quickly found out that atwo-dimensional graphmight be a bit easier to understandum so just to explain the results thatwe got here the red circles are channelsthat are only showing media to risecontent while the black circles show amuch broader variety of contentsand alsomake clear how this works these tinycircles are used to subscribe to thoseyoutube channels and to filter out asmuch noise as possible for this graph weare only usingshowing the users that are subscribed toa media to rise channeland in the left bottom we find a clusterof white nationalistic youtube channelsranging from commentators but also toeuropean fascist movements and just bylooking at the black dots we can alreadyfind channels that share a bigoverlap in audiencesand what’s also really interesting is inthat in the top right corner we found acluster of weightlifting channels theone with a big red dot is posting bothmedia to rise as well as weight liftingcontent and the user around thosechannels are subscribed to both thoseweightlifting channels as well as thechannels in the white nationalismcluster which really shows how thisparticular cluster serves as a gatewayto white naturalistic contentso the next step would be to move overto a more qualitative analysis and forthat we need to select the most relevantchannels for the palette list by variouswe just did the two spreadsheets but wealso have plans to optimize the tool forthat as welland from there on we can search withinthe transcripts of those relevantchannelsand through this we try to solve thatother problem that cameron mentionedearlier there’s just too much videomaterial on youtube and in many cases wejust need to find the relevant parts soin this example we’re specificallysearching for the wordswaukesha what are we doing nowwell we’re here in waukeshawisconsin in front of the courthouse asyou can see this isthe recent anti-white terrorist attack[Music]to play a song from my second album homeum to help promotethe new umtara and waukesha documentary by mediato riseandif i remember correctly on their teleand already we find many new connectionswhich we couldn’t directly see in anetwork of we see neo-nazi mike enochappearing in a video and apparentlythere’s also some sort ofsinger-songwriter actively promotingthis documentary and together withdesigner bennett lingard we’re workingon a narrative tracing functionalitywhich is a way to link different piecesof narrative together and in thisexample we are linking how these mediato rise documentary spreads in otherother communities such as the previousmentioned sing a songwriter one butwe’re also using a simple topic modelingimplementation to suggest new searchterms and that can be found in thetranscript that we found herebut in the future we also want toimplement suggestions based on on thenode positions in the network graph thatwe made so for instance with the earlieryou mentioned sing a songwriter the notethat is closest to her is actually theonline music festivalthat she performed in and another closenote is a musician that she hascollaborated with in the pastso i want to keep this tool freelyaccessible for academic research andjournalists but we’re also looking for amore commercial model to sustain thisand collaborating with the globaldisinformation index to explore futureusage of radiotube to limit the amountof advertisements on harmful youtubevideos and we think that could be thisreally interesting synergy between civilsociety users on the one hand and brandsafety on the other to subsidize thisand beyond searching for funding tofurther develop retitube we’re alsolooking for partners to collaborate withas well as better users in the nearfuture we’re also organizing workshopsin which you try to uncover harmfulyoutube communities such as the onei’ve been showing earlierand so feel free to reach out atinfo at reddittube.com and giving itback to betsythank you so much team excitingpresentation judges uh feel free to comeup and who would like to ask the firstquestionchris you want to go aheadyeah sure so um i thinkyou know in terms of success what doessuccess look like for the project whenit’s fully deployed you know what’sthere are a number of different usecases but how do you you know what’swhat’s the greatest possible applicationof the tool uh in your view is it on theplatform level is it aindependent ngo is it a governmentagency how are you thinking about thisum yeah i mean so for the kind of civilsociety part of this we’ve beenspecifically targeting mostlyjournalistic or academic researchers sogetting in the hands ofother researchers to help them doresearch on youtube is kind of our idealuse case and since that tool uh doesn’treally exist um and then in additionwe’ve been like eric mentioned at theend we’ve been trying to kind ofsubsidize that through some more uhcommercial partnerships and so i think imean ideally we want other researchersusing this to carry out this kind ofresearch on youtube sinceeric and i aren’t really confident beingarbiters of truth but we do want to kindof uh open up this um platform for uhgreater transparency i mean you can lookat like there was a recent letter from ithink 80 different fact-checkingorganizations um calling for like morethird-party research on youtube and moreplatform participation and so if we canencourage that that’s that’s awesomeyeah so quick quick follow-up umthinking throughyou know thethe tool it seems to me um the way youhave it kind of optimized right nowcould be repurposed i mean there are anumber of different given the searchparameters and things like that so doyou have any consideration of uhyou know kind of prevention of abuse ofthe tool for any nefarious purposes wethought through that aspect of ityeah so i meanthe most obvious is where we have beenvery careful in terms of like what datawe make public so we are um scrapingthrough users for instance to carry outthis discovery but we don’t want tohighlight any uh like users that are notuh already posting content and so we’refocusing on kind of content creatorsthat are already public uh as well umany kind of um larger scale users uh werequireum like you know authentication andaccounts so we can kind ofkeep an eye on like what the tool isbeing used for so basically the higherprofile channels we’re going to leaveopen publicly but for any sort of morespecific narrative tracing work uh thatyou know might be used to track activistorganizations things like that like thatthat will require um authentication andyou know participation of the team yeahgot it i have a question about umreally interestingwork and i i do think it’s important toshed more light on youtube for thereasons you’ve highlightedumwhat level ofdata access would you need to to operatethis tool at scale and if youtube becameawareof your activity is there anyis there any dependency onlevels of data access that could changeyes so um ideally we would like to havebigger access to the youtube api umright now we’re using a combination oftechniquesincluding some scraping as well umbut yeah i mean yeahthere is like we yeah we would very muchlike to collaborate with youtube on thisum actually because we do think it’sit’s very essential to do thatyeah and i will say like for the morecommercial work so some of the workwe’re doing with ad transparency umthat’s with a group who has kind ofactive uh active relationships withgoogle um so for any of of that work wedefinitely have to be more above boardin terms of how we do the datacollection um however i think there wasa recent decision that we were reallyexcited about uh in the case of acompany that was scraping linkedinpublicly and that was decided that thatthis kind of scraping of public data uhis allowed and um should be allowed tohappen and so we’re hoping that umyou know the the greater kind ofpressure on these platforms encouragesuh allowing more scraping andtransparency effortsamandaum a quick question maybe that follows alittle bit more um to chris’s how do youimagineum evolving i’m sure you’ve got acatalog of terms that you’re using andbuilding as you’re you know discoveringmore abusive and harmful content and umas we all know there is more and more umthere are more and more tactics to evadethat type of tracking and detection soi mean obviouslysome information is better than none atall but like how are you um envisioningstrategically just staying on top of thekind of semantic aspect of um you knowgetting as muchsignal versus noiseas possibleyeah i mean i think the theeasiest part of the i mean easiestanswer to that is so for like channeland content discovery we actually use alot of um other tactics that aren’treliant on transcripts or speech orterms like that so for instance we lookat shared audiences to determine sharedcontent rather than um the actual youknow rather than searching for specificwords um which helps discover thiscontent without that kind of like umsemanticevasion i guessas well we we kind of leverage likeother access to other platforms wheresome of this organizing happens as waystouh better track these communities aswell um and i will say that’s a toughproblem and i think part of the solutioncomes from the users to the tool and sowe can do our best to you know increasetransparency provide this data publiclyand give people help for signals interms of like uh other channels to checkout or other things to search for um butin the end it’s also up to kind of theresearchers using the tool to decidewhat is important for them to toresearch and we want to leave that opentokind of more active civil societymembers yeahfantastic that’s time thank you team foryour presentation we really appreciateit and now we’re going to move on to ourthird teamranking digital rights is a program atnew america focused on corporateaccountability for human rights in thedigital agefor this information prize competitionthe team has focused their project onmaking the online ad economy moretransparent and accountablepresenting their work today is nataliemarichal zach rogoff and alex rochefortover to you teamthank you betsy let me go ahead andshare my screen hereto fix information disorder we must fixonline advertising i’m natalie marichalpolicy director at ranking digitalrightsover the past two months my team and ihave been developing model ad policiesthat incorporate 10 essential practicesthat companies must adopt tosignificantly reduce the spread andthreat of this informationtoday we’re going to show you why wemust demand more transparency andaccountability from big techthe elements of our model policies andhow we built themand why starting with sound ad contentand targeting policies will lead tobetter outcomes for free speechmore effective content moderationtechnologies and greater integrity forour information environment as a wholeso why focus on adsthe pathologies of our global digitalecosystem are intertwined withadvertising supported platform businessmodels harms are in the ads themselvesin the discrimination that ad targetingenables and in the incentives theycreate for platforms algorithmicrecommendation systemsthey also provide revenue fordisinformation for profit outlets oftenwithout the advertiser’s knowledgethis image comes from our 2020 it’s thebusiness model report where we firstrecommended a focus on ads as the imageillustrates we’ve been saying that tocombat disinformation we must focus noton the speech itself but we must goupstream and look at the targeted adbusiness model and demand that companiesimplement ad content and targetingpolicies that anticipate and preventharms by adhering to human rights basedstandards and to companies ownobligations to respect and promote humanrightsright now online advertising is drivenby the logics of automation scalabilityand increasing profit margins there arehardly any safeguards to stop purveyorsof disinformation to take advantageour goal with this project is tointroduce transparency and strengthenaccountability to force companies tocreate and enforce these safeguardssince 2015 rdr has celtics set the goldstandard for corporate accountabilityfor human rights in the digital ageour research methodology evaluatescompanies policies and practices on 58indicators and three categoriesgovernance freedom of expression andinformation and privacyin 2020 we added new indicators andtargeted ads and algorithmic systemsnotably those used to curate andmoderate contentwe know that this is really central tohow big tech operates but so far ourresearch shows that companies are veryopaque about thisto develop model policies we need tounderstand first the harms that we canattribute to adssecond how those harms might bemitigated through better practice andthird what transparency we canreasonably expect from companieswe identified 10 ways to determinewhether a company is being transparentenough about its ad systems the ideabeing that this transparency is thefirst step toward accountability andevidence-based policy makingthen we applied these ten indicators ofsix existing ones from our establishedmethodology to meta and twitter’s adbusinesseswe looked at things like whether therules for ad content and targeting arepublic how the company actually enforcesthese rules respect for data privacy andwhether the company does prior duediligence to make sure its products willcause or contribute to human rightsviolationsthe results aren’t great but theimportant thing is that the indicatorsthemselves provide a roadmap forcompanies can create advocacy targetsfor civil society and a framework forpolicymakers as they consider regulatoryinterventionswe also went beyond our establishedroadmap approach and created modelpolicies that would earn 100 score inour assessmentif the research indicators provide aroadmap these model policies are thedestinationif you choose to invest in this projectwe’ll extend our research to morecompanies like video hosting sites suchas youtube and ad exchanges theintermediaries that facilitate adauctions for websites all over theinternetwe’ll engage with companies directly onthese issues and publish our report andpublic data setcrucially we’ll produce concreterecommendations for companies and policymakers which will form the basis forcivil society campaigns to bring ourrecommendations to lifei’ll now turn it over to my colleaguesto share some examples of the problemswe want to address and highlights fromour research findingshi i’m zach rogoff and i’m a researchmanager at ranking digital rightsfor me one of the most infuriatingthings about working in the platformaccountability field is how oftencompanies seem to be taken completely bysurprise by problems that basic duediligence would have helped themanticipatefor example in 2020 people were able toadvertise sales of human organs ontwitterand in 2021 the tech transparencyproject using only facebook’s built-insearch found many groups advertisinghuman smuggling services on the platformhuman rights impact assessments are amethodical way of thinking through theworst case scenarios for a company’sproducts they are foundational tobusiness and human rights yet across theboard companies are failing to do themmuch less take steps to mitigate riskswith the incentives of the surveillanceadvertising business model being as theyare this essentially means thatadvertisers can do whatever they wantuntil the platforms are publiclyembarrassed enough to stop themuntil the fair housing act of 1968 itwas common for mortgage lenders topurposefully exclude people of colorfrom some neighborhoods through apractice known as redliningredlining is partly responsible for thesegregation in this demographic map ofdetroitunfortunately redlining is echoed intoday’s digital redlining which canhappen when ad targeting parametersexclude people belonging to a protectedclass such as race age or genderthe point of targeted advertising is toshow different ads to different peoplebased on who they are and what they doonline sometimes that’s benign or evenuseful but it can also constituteillegal discrimination when it keepscertain people from seeing ads for jobscredit or housingdigital redlining and ad targeting canhappen two different waysone is for an advertiser to explicitlysay for example do not show my housingad to black peoplethe other way is for the platform to useopaque targeting optimization algorithmsthat even its own engineers do not fullyunderstand to further refine the adsaudience beyond what the advertiserspecified so that it performs betterit’s pretty clear that the systems forenforcing platforms ad policies need tobe improved but perfect enforcement isunlikely we need a way to audit the fullset of ads that run on platforms to findads that are against the rules and keeppressuring companies to improve theirenforcement systemsthe best way to do that is through apublic ad transparency database alsoknown as an ad librarywithout comprehensive ad transparencydatabases it will remain time consumingand expensive for third parties toverify that platforms are enforcingtheir own policies and policy withoutenforcement is just public relationsthanks heck my name is alex rocheforti’m a policy fellow at ranking digitalrights and a phd candidate in emergingmedia studies at boston universitylet’s start a review with a review ofour top line findings comparing meta andtwitterfirst of all both companies receivefailing scores meta does achieve aslightly better result than twitter andthis is due primarily to twitter’sfailure to disclose anything acrossseveral indicatorsthis is a particularly interestingfinding in light of our big techscorecard which places twitter at thetop of all other digital platforms itscores particularly well in categoriesrelated to freedom of expression as itrelates to advertising however itsdisclosures leave a lot to be desiredlet’s take a deeper dive into a fewspecific areasas zach explained whether or not acompany conducts impact assessments forits targeted advertising practices canhave important consequences for on andoffline harmsand the disappointing conclusion is thatneither meta nor twitter does well onthis issuemeta score is 21 here because itperformed a limited impact assessment onits targeted advertising systems as partof the civil rights audit it agreed toafter settling a case with the fairhousing alliancealong with the naacp the group had suedmeta for allowing advertisers topurposely exclude black people fromseeing housing ads which is illegalthere is no evidence that twitter hasever done this kind of due diligencethis is an example of how we would writea full model policy for impactassessments of targeted ads obviouslythis is a lot of text too much to walkthrough right now instead i want to justfocus attention on how this model policyincorporates the various indicatorelements the yellow text bubbles pointto specific aspects of the policy thatwe evaluateour next example asks whether or not acompany clearly discloses that does notperform tar algorithmic targetingoptimization after an advertiser setstargeting parameters to avoiddiscriminationto receive full credit on this element acompany needsonly to add an additional sentenceclarifying that do not permit thispractice of course as with all of ourindicators a company also needs to dowhat it sayslooks like a mismatching slidethere butwe do also evaluate umad transparency databases uh we’ll getinto the specific aspects of what wewould expect such a database to discloselike the scope of the library targetingparameters using that deployment thecost of an ad and moreonce again meta receives a higher scorethan twitter with 36 to twitter zerothis is because twitter shut down its addatabase after it stopped runningpolitical and social issue ads in 2019.herei’m just waiting for the slide changeand here we have a model policy for thead library you’ll note that we have ascreenshot of meta’s library and an xover twitter is now defunct databasehere we want to make the point that notall policy disclosures need to belengthy although implementing thecomplete set of ad library disclosuresfor a full score would demand additionaltechnical and operational steps thepolicy itself can and should bestraightforwardnatalie you’re on mutemy lord you’d think i’d love learnedabout this over the past two years uhourinvestment in our project would besupporting a team with a proven trackrecord of impact our work influencescompany policies shareholder resolutionsnational legislation and the work ofintergovernmental institutionscivil society groups around the worldadapt our methodology to research thecompanies and issues most relevant tothem and they do advocacy campaignsbased on the resultsby funding us you’ll send a strongsignal to advertisers to policy makersto advocates and most importantly tocompanies that the era of lawlessnessand online advertising has got to endyou don’t have to support a ban onsurveillance ads to support basictransparency and accountabilitystandards like the ones we’re proposingfor years now silicon valley leadershave been telling us that before we knowit their ai systems will be good enoughto rid the internet of this info andother harmful content and we just needto trust them to nerd harder that’snaive techno solutionism at bestmachine learning has a role to play buttime is coming not enough thank you backto betsy very muchthanks team we really appreciate it umand last but not least we’d like tointroduce our final team the observatoryon social media based at indianauniversity the observatory on socialmedia unites data scientists andjournalists to study the role of mediaand technology in society and buildtools to counter disinformation andmanipulation online today the team isshowcasing an open source public tool toidentify misinformation super spreadersaccounts that amplify bad information atlarge scale representing the team todayare phil menzer and matt diverna over toyou phil and matt hey beth see realquick are we gonna be able to ask oh i’msorry i’m so sorry questions thank youwe’ll go to questions first then i willhave already done my introduction my badchris over to youyeah i just i think the kind of the topline question i think you answered inthe last slide and there was a questioninthe umin the q a box about incentivizingparticipation so can you speak a littlebit more about your model and you know iwas thinking through your elements andhow you came up with them and how wouldthey know but can you speak a little bitto the foundation you’ve laid over thelast several years in terms ofpromulgating a set of expectations thatcompanies might be able to meetsure thing so the beauty of our model isthat it doesn’t require companies to uhto participate we do the evaluationbased on publicly available informationand then we publish the resultsso the companies that we evaluate in thebig tech scorecard by and large doengage with us throughout the researchcyclethey don’t always agree with thestandards that we setbut they accept that they that they havenormative value and that there’s a lotof buy-in from the community there andso the idea is to uh set a standard thatthat is grounded in uh in expert uhconsensus and uh tobuild uh build pressure on the companiesfrom all directions uh to move in in thedirection that that our roadmap sets forthemokay and quick follow-up as i waslistening it was clear to me that you’reobviously measuring and you’re you’reevaluating the platforms themselves areyou also evaluating the companies thatare actually advertising and how theydecide to engage on the platformsbecause i think that’s a usefulindicator tooin in corporate space sure so that’ssomething that that we would do as partof the next phase of the project but itwas not part of this prototypeuh i guess a question to sort of buildon that iam just wanting toi think i understand that there areseveral points of pressure that couldcome fromsimply compiling this publicly availabledata and making it more accessible whichi think is a noble effort but what isthe desired effect really here is itto inform legislation so that there ismore oversight of this is it topressure the com the platforms to createthis ads database is it toum inspire you knowcompanies to change ads that are alreadyout in the wild likethere are lots of different areas forpressure and i’m just wondering if ifthere’s a priority set there for you allthat you’ve thought through or thatresearch helps informgreat question realistically withadvertising the the goal is to influencelegislation or regulation if companieswant to get ahead of of that process andstart voluntarily meeting meeting thesestandards that’s wonderful but i dothink that legislation or ftc rolemaking will be requiredthank youjust to clarify the when you talk aboutthe ad database you might have said thisin the presentation i missed it umyou’re also talking about umsome kind of transparency of the reachand the sort of differentumdemographic dimensions of the audiencethat is reached for the ad that’s partof what you’re calling for ideally wouldbe made transparentuh well we were specifically talkingabout the targeting parameters so we canknow who the uh who the advertiser wasintending to reachbut yes we would also like to see whohas been reached so that you can seewhat the delta is between those twothings because if there is a differencebetween the intended audience and the umde facto audience that strongly suggeststhat the platform introduced some levelof discriminatory targeting through theoptimization algorithms that zach’stalking about for examplewell it may be that or it may be thatthere’sum underlying correlations betweenum attributes of the population and thethe targeting dimensionsit could and then that may or may notactually constitute uh illegaldiscrimination on under different civilrights statutes um the devil is reallyin the details there but the point ofthat this the point of disclosing thisinformation is to be able to pinpointwhere there might be discrimination toallow for further investigation and thenremedy of the problem so just to beclear the what you just proposed to usincludes both the in the the targetingum intention and theactual reach you that’s rightokay thank youwe have time for one more quick questionif there is anyif not i’ll read one from the chat whatincentivizes social media companies todo thispublic pressure reputational harm andthe fear of regulationfantastic that’s a great note on whichto end thank you and again sorry foralmost skipping your q a but we’ve gotlots of great people who fixed it for usum so now uh we will turn over to theobservatory on social media phil andmatt over to you this time for realthank you betsyhi everybody i’m phil manser i’m thedirector of the observatory on socialmediaor awesome as we like to call ourselvesand i’m here with matt durenaa knight fellow at awesome and phdstudent and we want to tell you aboutmiss amplifiersa public tool that we would like tobuild to identify hidden superamplifiers of misinformationwell that’s a mouthful but of coursewe’ve all heard about misinformationsuper spreadersthis was initially highlighted in apaper in 2019that showed that 0.1of users so a very tiny fraction wereresponsible for sharing 80of fake newsand this has been talked about a lotmany of you may have seen the reportmore recently from the center forcountering digital hate on thedisinformation dozenshowing uh in this case in the contextof anti-vax misinformation that again avery small number justa dozen of very influential accounts areresponsible fora great portion ofanti-vaccine misinformationwe’ve studied this problem ourselvesin particular among others in thecontext ofmisinformation aboutcovet vaccines we have a platform calledcovaxy that matt has co-developedwhere we track data aboutcobit vaccineespecially from low credibility sourcesand we look athow that can be associated with realworld outcomes for example decreases invaccination rates at the state at countylevel and we just found some veryinteresting results were published a fewdays agonow in looking at the data both fromtwitter and facebookwe look at where is this informationthis misinformation coming fromsorry for the complex plot but the takehome message here if we look at the leftis thatthe original posters um you know thereis there’sthere’s a great variety of accounts thatpost the originalmisinformation but if you look at mostof the postsuh that come in terms of reshares orretweets they come from a much smallerset of users so uh these these these umyou know most of the retweets originatefrom a small number of accounts again wecall these super spreaders and not onlythat but again on both platforms itturns out that these tend to benot unknown accounts but elitewell-known influentials often verifiedaccountswho are theywell on twitter for example the numberone spreader of misinformation recentlyhas been children’s health defensei’m sure many of you have heard about itit is disheartening to see that it getsmoremore rich more posting more visibilitythanreliable sources like cdc and wh sohere’s an example of a tweetandyou see that in this particular case thetweet has been retweeted by robert fkennedy who runs this organizationandso if you we are attributing uh all theanalysis based on super spatters todayare attributing this information to theto the origin to children’s healthdefense butin this case a lot of people may havebeen exposed to that information throughan intermediary like this account thatactually happens to have four times asmany followers as the organizationaccountso it could be that there is a veryimportant and unseen amplificationeffect that you don’t observe if youjust look at the twitter data as we’vebeen doing until nowso that’s what we want to do we want tohelp identify hidden super amplifiers ofmisinformationand quantify quantify and visualizetheir impact these are these arethese are accounts that normally are notobserved in current analysis of superspreadersit turns out that this actually is aproblem that poses a very serioustechnical challenge now uhjust to give you a sense of that let meshow you a very simple example let’s saythere is a piece of misinformation thatis shared and a bunch of accounts uhretweetedthis is this is what the data that weget we and any other researchers todayget from twitter all of the retweetsuh are retweeting appear to beretweeting the original poster so thatwould be you know if this piece ofmisinformation goes viral that that’sthat’s the account that would be labeledas super spreaders however imagine thatwe could reconstruct the actualdiffusion cascade and see who’sretweeting whom we might notice as inthis example that there is anintermediate accountthat and that is really the superamplifier the original account maybe wasretreated only a couple of times butmaybe at a later time a very influentialaccount steps in retweets and now mostof the people are actually exposed tothat misinformation through thisintermediaryso this requires reconstructing thediffusion cascade umthis is not possible using directly datafrom available from twitter even lessfor other platforms but what we’ve beenworking on is efficient approximationalgorithmsto reconstruct these cascades with withhigh accuracy andto illustrate how we do that in ourearly prototypei’m gonna uh letmatt uh show you a live demo of of ourcurrent prototype that we just built inthe last few weeksuh matthi everybody umi’m really excited to show you guys mysample fighter today soas phil mentioned the the purpose ofthis tool is to allow people touncover hidden amplifiers misinformationon twitter so we imagine a website thatwill looksomething like this very simple to usea user can come to the website andsearch for hashtag a keyword a linkreally anything you can think aboutsearching for on twitter we would wewould let them sort of search formisinformation campaigns related to thatsearch for today for the prototype i’mgoing to show sort of a random a randomsearch so you can imagine a user wouldcome and click this search loadcredibility for links buttonand you can see that misamplifiercreates this really cool3d network visualizationwhere the nodes here are representativeof users and the edges arerepresentative of retweets um soholistically what we’re getting in thisin this example is a sort of randomsample of misinformation um that we haveunder the hood already reconstructed forusers so this is really the power of thetool is that we’re solving this thistechnical problem that phil’s describingbefore um and sort of removing it fromthe problem for anybody who wants tolook at the sort of better data that youcan’t really get from twitterumso uh on the left you’ll see we also getsome additional information here so wehave the the top 10 originators so theseare people who are posting theinformation and sort of uh who you wouldconsider to be the problem if you’rejust looking at the raw data fromtwitter and then we also have the top 10amplifiers so considering the all ofthese reconstructed cascades um you knowwho do we think is actually the largestproblemas you can see sort of flipping throughbetween these liststhe colors of these nodes are changingsoto sort of explain what that meansessentially the red nodes are basicallywhatever userthe user is in whatever list we’veselected and then the yellow nodes areanybody who sort of immediatelyretweeted that problem user umand the blue nodes or anybody else inthe cascade something i haven’t uhexplained yet is the sort of size of thenode so this includes some informationas well which essentially is justshowing us the total number of retweetsthat a user is getting across all thedifferent cascades so you can see we’veselected the originators here so this isthe worst people according to sort ofthe naive data from twitter and we cansee in this cluster specifically we’remissing sort of the biggest nodes thatare at the center of this misinformationclusterthat are really sort of responsible forspreading thatthe content so with the tool we can justsort of just select the amplifiersand begin to begin to explore a littlebit so you can see here in this clusterwe have kim guilfoyle and donald trumpjr that’s maybe not surprising thatthey’re sort of close togetherlinda seller is this account that maybei’m not aware of who that iscernovich and coulter so again the wholepoint of the tool is to explore a littlebit so i mentioned this account that idon’t know um but what we can do withthe tool is double click the nodeand this will sort of take us directlyto their twitter profile we can see okaylinda seller thinks that president trumpwas the best president ever um we cansee that she has 347 000 followers soyou can start to get an idea how youknow if this person isn’t actuallyposting original misinformation thatthey could sort of use that influence toamplify uh you know that misinformationwithout it being seenand i can look at one other cluster hereso um if we look at this one we see wehave the dem coalition account um grantstern account so this this looks maybe alittle bit more liberal um and again wecan do the same thing right so if wewanted to double click on this node wecan check out grant sternum and we can see thathe’s the executive editor of occupydemocrats which is a low credibilitysource on the leftum and and hopefully you know i’ll pausehere for time but hopefully we can kindof get a feeling for sort of the powerof this tool and its simplicity is justthat we’re sort of abstracting away thisdifficult technical problem of doingthis reconstruction and giving peoplebetter data whether they are aa researcher or a journalist or you knowreally anybody who’s interested inlooking for for amplifiers uh and withthat i’ll pass it back to phil to wrapthings upthank you very much mattuh so just to wrap things up uh why doesthis stuff matters well going going tothe to the aspen uh reportwe are addressing especially the issueissues of accountability accountabilitynorth but also accountability aboutsuper spreaders this will allow socialmedia platforms to hold the mostproblematic actors accountable andgovernment and independent organizationsto hold the platforms accountable forfor that moderation and also uh in inthe area of healthy digital discourse wecan quantify the impact of these hiddensuper spreaders andthis could supportmoderation policies that maximizereduction in harmful information withminimal moderationand of course it will also help raiseawareness about the role ofmisinformation amplifiers so why ask whyawesome well we are a research centerthat studies online misinformationspread and we’ve developed a lot ofstate-of-the-art tools to identifyinauthentic malicious or harmful i’msorry your time is out i’m just uh yeah10 secondsuh and these tools are used widely byresearchers journalists civil societyorganizations so we are we’re sure thatthey make an impact okay thank you somuch and we’re happy to take yourquestionsfantastic questions from the judgesamanda would you like to go firstsureumthis is fascinating i guess you sort ofspoke to this toward the end but again iwas just coming back to like thethe ultimate goal and is it that you’rehoping to compel the platformsthemselves to take actionon um another sort of subset ofproblematic users or is it really morethe public awareness piece of it i’mjust wondering about the platforms iimagine they probably do have this dataalready and don’t act on it and so kindof wondering what bet you’re going tomake there in terms of impactso if you look at our past research andtools you know we’ve developed tools forexample that have shed light on theissue of social bots and uh and theissue of of coordinated inauthenticcampaigns and these once the research isout there and the public becomes awareof those issuesthen platforms usually respond and we’veseen that recently for example twitterhas become better at moderating theaction of inauthentic accounts and sohas facebook so we think it really comesfrom all perspectives we want to makethese tools available to the public wewant to make them available toresearchers so that they can more easilyanalyze data about thesediffusion cascades data that right nowis really hard to to to get because theexisting algorithms are really uh veryvery complex it’s very difficult to doanything like this in reasonable timeuh getting for example follower data uhso so this will support research it willsupport end users but it will also ithink help put pressure on social mediaplatforms toum you know to pay attention to who isamplifying misinformation and this isespecially relevant these days as wehear talks about big changes andpossibly reinstatinguh some accounts who have spread a lotof harmful content in the past sowe think that can have a big impactthank youthank youit seems um just to maybe clarify theintent of the toolumitseems more general purpose than evenmisamplification or amplification ofmisinformation of justgiven a topic of interest or a hashtagsome kind of queryto findinfluential nodes in the diffusion rightregardless of whether it’smisinformation or not right it’s just uhyes deb you’re absolutely right ofcourse here we wanted to focus on the onthe dc formation portion which we thinkis particularly important but yes uh thetool could be a generaltool in terms of detectingmost influential accounts whether or notthey originate the information so you’re100 correctyeah i was picking up on that somewhatso kind of two questions first is i meanthere are other i think commercialapplications for the tool you knowcompanies that get brigaded or you knowa wayfarer-like eventum you know have you have you thoughtthrough first in these commercialapplications that’s one way toadditionally kind of find somesome additional resourcesand i think the second isis there any abilityto settriggers or thresholds azure monitoringor as you’re watching something emergethat that you know one of the thingsthat we always look for is the reach theengagement is this a significant uhlevel of amplification and at what pointso are there you know do you have thatsort of quantification of what severitylooks likeso i’ll i’ll start and then i’ll letmatt at any thoughts uh so on the issueabout umcommercialization uh normally we aredriven byby research and all of our tools areopen source publicly availableum because we want to sort of maximizethe applications byyou know in the research communitygenerally civil society organizationsso so far we we haven’t[Music]thought about that it’s possible thatthere are commercial applications thatusuallysome organizations may come out and makecommercial versions but we think firstof all it’s important to get theresearch out and toraise awareness and tobuild a tool that is easy to use so thatanybody can sort of observe andvisualize what it means to to amplifybecause we hear a lot about terms oflike amplificationbut uh but there is nothing out there toreally kind of make it easy for peopleto to observe thatumso that’s an open question maybe butthat’s not our immediate goalin terms of your second question umyou could think of lots and lots andlots of bells and whistleswe’ve thought about many and it’s it’sbeen hard to focus this demo thisinitial prototype on on something youknow narrow enough that we coulddescribe in five minutes umbut yes definitely if we are allowed tocontinue the development uh we have lotsof ideas and lots of thoughts uh thereare different ways that you couldimplement this uhusing live data from twitter using datathat is available that other researcherscan submit uh people can run of coursethe algorithm themselves if they want toumthere are ways to quantify i mean mostof the work the technical work under thehood that matt and and uh rasheed andthe rest of the team have beendeveloping is exactly developingefficientalgorithms to quantifyuhand and therefore be able to rank andidentify the the worst offenders sothere is an important work in the areaof of metrics there and in fact we’reworking on a paper right now uh matt doyou wanna add that i’m sorry i’m sorryphil we are actually at five minutes sounfortunately we don’t have time for anyfurther points but thank you so much tothe observatory for your presentationsand thanks to all four teams uh what afantastic show we really appreciate itso now the judges are going to return tothe green room to deliberate and in theinterim i’d like to hand the floor overto my colleagues vivian schiller andryan merkley to provide an update on thework of the commission of on informationdisorder if time allows our deputydirector my sister will moderatequestions over to you vivian and ryanand we’ll be back soongreat thank you so much and um thank youeverybody it’s really thrilling to seethese really creative uhsolutions to a lot of the challengesthatthe commission laid out in the reportand we’re so thrilled to have three ofour 16 commissioners here with us um i’mmostly going to the more interesting uhstuff we’re going to talk about is sortof what has happened since the report umwas released but let me just say a fewwords for those of uh who are watching alittle bit about the commission and thereport and then i’m going to turn itover to my colleague ryan to sharewhat’s going on so you will recall uh ormaybe or learning for the first timethat um aspen digital again we’re aprogram of the aspen institute we’reproud that the uh tech policy hub ispart of our team that we uh curated agroup of uh of commissioners from allwalks of life three of whom you see hereum that bring different perspectiveslived experiences expertiseand uh they spent about six seven monthsumfirst uhgetting versed in all the aspects ofinformation disorder that they weren’talready familiar with uh most of themwere are super duper expert on someparts but maybe not on all parts lookingat the various solutions and proposalsthat are out there from the brilliantresearchers and policy makers both inthe u.s and around the world they thencame up with three priorities for wherethey were going to focus uh thoserecommendations which were about umaddressing harms umabout transparencyand aboutoh i used to just have this roll off mytongue ryan what’s the third priorityhelp metrust trust thank you thank you the mostimportant one of allum and then um and then narrowed downyou know was very hard work um a seriesof uh recommendations uh that thatcomprise uh most of the report that aredirected at the government uh this ismostly u.s focused um at the state andfederal level to um tech companiesto private industry outside of the techindustry uh and also tocivil society organizations and franklyto populations um wewere really delighted with the attentionum that we got and the embrace of thoserecommendations which immediately led toa lot of coverage and a ringingendorsement from the washington postuh but you know that was not the end ofthe line uh we did let our commissionersum take a break after that but we havecontinued to pursueall of these recommendations umuh with a with a with us with a goal ofuh gaining traction and to talk aboutupdates and progress since then i’mgoing to turn it over to ryan merkleywho is the managing director of aspendigital and was theleadoverseeing the commission on informationdisorder for aspen digital ryan over toyouwell thanks vivian um and thankseverybody for your presentations and forthose of you who are watching uh wereally appreciate you being here uh anduh they’re just really some great workum i’m gonna give you an update on someof uh the work that we’ve been doingsince the release of the report so youknow following the release of thecommission’s recommendations we’vecontinued to engage in a set ofconversationspretty widelyandyou know as you’ve likely all had a lookat the report those 15 recommendationsreach broadly across a variety of areasand so we’ve had many different kinds ofconversations we’ve spoken with folksfrom the biden administration we’vespoken with members of congress withpeers and colleagues working on thedisinformation challenges directlywith corporations and funders and civilsociety groups that are interested indifferent aspects of the report whetherit’s restoring faith in the mediaaddressing issues around norms trust asvivian was saying or transparencya number of the commissioners have alsobeen speaking about their work uh and inparticular around the aspects thatthey’re most interested in um in earlyjune we’re actually going to get ourco-chairs uh kris krebs katie couric andrashad robinson back together at the rsasecurity conference in san franciscowhere they’ll be delivering the closingkeynote for that eventum so i want to share with you a coupleof highlights uh that uh have happenedsince the release of the report just togive you a bit of a window into thekinds of things that we’ve been doingto keep that work going and to buildmomentum around some of those thingsthat we highlightedin the recommendationsso in january we hosted a congressionalbriefing for interested members ofcongress and their staff to walk themthrough the recommendations to heartheir questionswe’ve also as i said at the beginningreached out to the biden administrationto provide them with information andensure that the report’s recommendationshad the widest possible reach about athird of the recommendations call foraction from the government so we wantedto make sure that they had anunderstanding of what was in that reportand an opportunity to ask us questionsa number of recommendations relate tolocal journalism and in february throughanother project that aspen digital wasworking on with the lend fest institutea foundation they’re a foundation that’sfocused on saving local news andbuilding more sustainable mediaenvironmentswe hosted uh with them an event inoceanside california where we talkedabout the challenges and theopportunities facing the media ecosystemthe recommendations that were in thereport were actually referenced often inthat in those conversations includingthe need for greater diversity uh to bereflected in newsrooms and theirleadership uh and the role that avibrant local news ecosystem can play inrebuilding trust in a successful andhealthy democracyi’ll talk a little bit about our cybersummit so the commission on informationdisorder was actually originally an ideathat came out ofthe aspen digitals cyber group whichfocuses on cyber security issuesit identified disinformation as anemerging security threat which ledultimately to the commission’s work nowat the cyber group’s fall meeting inwashington dc the topic ofdisinformation including its applicationin the war in ukraine came up often andwe’re going to keep looking for waysthat that cyber security work and ourfocus on disinformation can intersectand cross-pollinatea few weeks ago at stanford presidentobama spokeon the topic of disinformation an areathat he’s been increasingly focused onand in advance of that meeting hepublished a reading list of articles andreports i will say we were absolutelydelighted to see that the commission’sreport was at the top of his list andperhaps even more encouraging was thatmany of his solutions were aligned withrecommendations that were outlined inthe report so we’re happy to seethat having further reach and interesti guess lastly is a thing that’s comingsoon the commission’s composition andits focus were really ideal fordeveloping insights and recommendationsin a u.s context um but we knew thatthat was limited we knew we had thecapacity to deliver against that thatsort of more narrow focus but we alsoknewthat we wanted to engage more broadlyin a global conversation about some ofthose issues after the commissionconcluded its reportso we’re actually beginning that workthis summer and in planning right nowwith an initial series of closed-doormeetings in some european cities in junewe’ll be in london berlin and brusselswe’ll be looking uh and hosting a seriesof exploratory meetings to identify thesort of next round of topics that wewant to explore but also looking forcommon threads that come out of thereport and its recommendations to seewhat of those recommendations might alsohold true in other contextsnowwhile we think the commission uhstructure was pretty successful we’llprobably look for alternatives thatallow us to have an even larger dialogueand there’ll be more to come on that andwe’ll be excited to share that as thingsdevelopuh so that’s uh you know some of thethings that we’ve been up to vivianmaybe i’ll hand it back to you and andmai if you want to take questions if wehave time we’re happy to do itthat pretty much covers where we are uhi don’t know if we have time my for anyquestions but we’re happy to take themif there are anyyes we do have time as the judgesdeliberate so a reminder to the theaudience members to use the q a boxbelow if you have any questions for ryanand vivian and thank you so much uh ryanand vivian for for giving us an updateuh on what the commission has been doingever since the report was releasedum the first question we have is do anyof the recommendations that uh thereporttalks about or any of your subsequentconversations uh touch on theadvertising supply chainuh so this audience member states somany publishers use intermediary inintermediaries to sell ad yet they don’tseem to take any responsibility forthoseuh yeah there’s there’s two that i thinkare the sort of obvious connections tothat um oneuh which is a very direct connection toactually three i guess so one is aroundadvertising transparency so one of therecommendations asks for a database uhand a tracking of all ad transparencywhich is a sort of prerequisite for thekind of conversation that i think thequestioner is leading us to which isto answer the question of who’s involvedand who’s being targeted you need toactually have that information and sothat’s sort of one fundamental piecethat contributes to being able to answerthat um secondly in the recommendationaround section 230 uh which is aboutimmunity for the platforms againstposted content one of the commission’srecommendations and also i would noteformer president obama’s recommendationsum is that the umuh is that advertising no longer be umuh exempt from um immunity on theplatforms meaning that there would be asort of higher level of accountabilityfor the content that is postedwhich we think would lead to some somegreater accountability forall the players involved in theadvertising industry i guess the lastone is the recommendation around normswhich talks about the sort of non-lawbased ways that we think about how werelate uh and one of those key norms ina ways the way that consumers andcivil society can apply pressure is toforce companies to be more accountablefor where they spend their advertisingand how they target it be it whetherthey’re targeting it to groups in waysthat are not appropriate or whetherthey’re working on platforms that don’tbehave appropriately i think thatcreates an opportunity for anaccountability amongst thosethose advertisers and advertisingintermediaries that facilitate that kindof spendingso exerting consumer and public pressurearound normsgreat thank you ryanuma question have you heard have you hadseen any response from the privatesector on on your recommendations in thereport to talk a little bit aboutengaging with government and civilsociety but anything from the privatesectori can take that one yeah we’re actuallydeep in there’s nothing really toannounce yet but we’re actually deep inconversation and when we think about theprivate sector um in the case ofthe commission on information disorderwe’re thinking about it in two partsthere is the tech industry the privatesector tech industry to whom a lot ofthese recommendations are quitespecifically targeted those that have umthat that provide user generated contentthat are driven by algorithmic decisionmaking and targeting and have the dataand there’s a number of differentrecommendations for them but there’salso you know reference and and ryanmentioned our norm’s recommendationabout um the uhthe rest of the private sector who arethe economic engine frankly for theplatforms um you know the fortune 50 100500 companies that uh both put out theirown content um that they generate butalso umuh put out you know what end up as adsthat again drive the the uh the businessmodels for these platforms and whatresponsibility they have in terms oftheir behaviors and actions um in in inmany forms across the board on theecosystem so we’re talking to them aboutthat as welli would just jump in and add to thatthat really there’s sort of at both endsof the spectrum there’s opportunity andthere’s risk uh the report uh kind ofhighlights right up front that um youknow increasingly we are seeingdisinformation targeting companiesdirectly so if your dominion votingmachines or if you’reyou know a pharmaceutical company uhyou’ve been the target of that kind ofdisinformation and then the opportunityis for leadership uh the edelman trustbarometer identified ceos as one of themost trusted groups i think that’s anopportunity for them to betruth-tellers both to their employeeswhere they have large bodies ofemployees or to the public about thingslikegetting a vaccine or you know otherissues that are of public interestwonderful thank you um i think we are attime now so i will hand it back to betsyto transition to the uh announcement ofthe winnerthanks so much for your patienceeveryone uh we’re excited to announcethat after a quick 15 minutes we do havea winner and so i’m gonna pass it overto our commissioner chris who will uh berevealing the winner chris over to youhey uh thanks betsy and first off justcongratulations to all the participantsuh all four projects were uh were reallyimpressive there’s a lot of thought ipersonally appreciated how you alltracked against uh the therecommendations of the commission and iyou know it felt a little you know felta little seen that the things that wehad recommended last year uh havemanifested in in some projects hereuh all that said there’s there’s got tobe a winner here right there’s there’sgot to be someone that takes home uh the75 000award and after as betsy mentioned thedeliberation we uh we picked agents ofinfluenceso congratulations to the team[Laughter]congratulations team uh we are veryexcited to award you um andcongratulations on your well-deservedvictory uh we’d like to take a moment tothank everyone who made this eventpossible uh first we’d like touh we’d like to thank craig newmarkphilanthropies and ex-ante an initiativeof smith futures uh for having uh givenus the opportunity to pull this togetherwe are extraordinarily appreciative toyou both for that fundingum to our contestants and judges allfour contestants did an absolutelyamazing job presenting and it was a veryvery difficult decision thank you somuch for taking the time to our judgesfor your evaluations thank you very muchfor being with us and for taking thetime to really uh think through thesequestions which are so important uh tomy sister meha aluwalia and maeve snedenof the aspen tech policy hub team for aflawless webinar other than me and forhelping to pull together such an amazingcontest and to vivian ryan diera and thecommission on information disorder forinspiring this contest and for helpingto select the semi-finalistsinformation disorder remains one of theworld’s great challenges and we’redelighted to have taken a small steptoday towards improving our approach forit please stay tuned for futurechallenge prize opportunities as we lookforward to taking small steps towardssolving big problems at the intersectionof technology in societythanks so much again and have awonderful day
On May 3, 2022, the Aspen Institute’s Tech Policy Hub and Commission on Information Disorder held the virtual pitch final event of the Information Disorder Prize Competition.
The prize competition, announced in November, sought to fund unique and innovative projects that make meaningful progress towards ending information disorder, in direct connection to one or more of the Aspen Institute’s Commission on Information Disorder’s 15 recommendations for government, industry, and civil society.
semi-finalist teams
Alterea, Inc.: “Agents of Influence,” a spy-themed media literacy video game for middle- and high-school students. Team members: Anahita Dalmia, Jasper McEvoy, Alex Walter
RadiTube: “RadiTube: Narrative Detection and Analysis for YouTube,” an automatic detection and tracing tool for harmful narratives spread through YouTube videos. Team members: Cameron Ballard, Bernhard Lenger, Erik Van Zummeren
Ranking Digital Rights at New America: “Treating Information Disorder by Making Online Ads Accountable,” a set of sample policies to make the online ad economy more transparent and accountable. Team members: Nathalie Maréchal, Anna Lee Nabors, Zak Rogoff, Alex Rochefort
Observatory on Social Media at Indiana University: “Detection Tool for Misinformation Superspreaders,” an open-source public tool to identify misinformation superspreaders. Team members: Filippo Menczer, Rachith Aiyappa, John Bryden, Matt DeVerna
Winner
Following the project presentations, judges Chris Krebs, Amanda Zamora, and Deb Roy of the Commission on Information Disorder selected Alterea, Inc as the winner of the $75,000 grand prize.
{"includes":[{"object":"taxonomy","value":"127"}],"excludes":[{"object":"page","value":"183669"},{"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":[]}
The threats posed by mis- and disinformation loomed large ahead of 2022. What are the risks that faced the elections and what can we do to counter them?
One year after the January 6 attack on the US Capitol, the State of Democracy Summit explored the most important questions facing the country at the moment.
January 20, 2022
THE BEST OF ASPEN DIGITAL, RIGHT IN YOUR INBOX.
Sign up for our email list to stay current on our work.
Facebook
Twitter
LinkedIn
Email
