Aspen Digital

In this high-stakes election year, it’s never been more important that voters have unimpeded access to reliable and truthful information about the candidates, the issues and the voting process. Amid an already fraught information ecosystem, artificial intelligence (AI) adds an additional alarming layer of risk.

That’s why Aspen Digital’s AI Elections Advisory Council created three risk checklists focused on areas where AI tools make it easier for bad actors to discourage and disinform: hyperlocal voter suppression, language-based influence operations, and deepfaked public figures.

The Advisory Council is a non-partisan group composed of civil society and technology leaders taking steps to build democratic resilience in the face of AI. The Council is chaired by Alondra Nelson, Klon Kitchen, and Kim Wyman. It is part of our ongoing AI Elections Initiative, which works to secure the US elections in November and beyond.

To help ground public discussion and leadership action in the facts underlying specific AI concerns this election year, these new AI Election Risk Checklists center on:

• Hyperlocal Voter Suppression: For years, bad actors have attempted to impede voting by spreading false information about when, where, and how to vote. AI tools can generate convincing content quickly, including personalized details and interactive exchanges that add credibility to false information. These tools make text message campaigns, interactive robocalls, and fake local news websites cheaper to run at scale.
• Language-Based Influence Operations: Artificial intelligence makes it easier to create content in any language, using automated translation tools. While these tools can be beneficial in many settings, in the wrong hands they can make spreading falsehoods easier, faster, and harder to detect.
• Deepfaked Public Figures: As artificial intelligence improves, it has become easier to create convincing images, audio, and video depicting public figures saying or doing something that they did not.

These risks are not inevitable. The checklists detail steps that election administrators, social media and messaging platforms, AI labs and companies, news media, advocates, and civil society groups should take to mitigate the harms that AI could pose to our democracy.

As Americans prepare to vote this November, we must not let fears of artificial intelligence deter us from participating in our democracy. Bad actors might try to keep us home this election, but they can only succeed if we let them. At Aspen Digital, we’re committed to partnering with those committed to election resilience to keep our democracy strong and to help ensure that every American has the information and access they need to cast their vote.