Email not displaying correctly? View it in your browser. |
AUGUST 5, 2024
FUTURE US NEWSLETTER |
Future of Security: The Future US recently convened the 4th policy lab on AI and Election Security, bringing together government officials, AI technologists, social media policy leaders (Snap, Meta, et al.), and civil society leaders to discuss deepfakes and the 2024 election.
- As part of our ongoing work, we are partnering to establish an AI and Election Security War Room and Tech Fusion Center with ISD.
- The Future US is teaming up with Represent US to develop a series of campaigns on public inoculation efforts on AI and Election Security.
|
|
Building on the first 3 Policy Labs, The 2024 Deepfake Dilemma Policy Lab, hosted by The Future US, Miles Taylor, brought together government officials, AI technologists, social media policy leaders (Snap, Meta, et al.), and civil society leaders for a threat update on deepfakes and the 2024 election. This included former tech advisors in the administrations of Joe Biden, Donald Trump, Barack Obama, and George W. Bush.
- The conclusion? Bad actors are already using AI-powered tools to undermine the election. The U.S. Intelligence Community assesses that AI-powered interference will get worse as we get closer to November — and Americans are largely unprepared for a possible deluge of deepfakes regarding the time, manner, and place of voting
- However, tech leaders were keen to emphasize that “deepfakes” are not the enemy. Digital twining and deepfakes will have beneficial use cases. But we need to help the public spot content meant to deceive voters, i.e. a “see something, say something” for election deepfakes that are designed to misdirect citizens about where/how/when to vote.
- The gathering was hosted by heavyweights — The Future US, Snapchat, Business for America, the Wilson Center, Represent.Us, German Marshall Fund, and Institute for Strategic Dialogue — with a commitment to reconvene again ahead of the election.
- Top Quote: “This isn’t a partisan issue. Very bad actors want to confuse American voters about their right to vote. And we need tech leaders working with public officials to protect the election from these deepfakes.” — Congressman Don Beyer
The Lab surfaced novel and specific threats to the election. Participants "war gamed" how deepfake tech might be used in the lead-up to and after the 2024 vote. They assessed known vulnerabilities but also surfaced NEW potential threat vectors. A few of the categories explored included:
- We are likely to see “leak fakes” this cycle, which include purported leaks of non-public information about candidates or elected leaders that are highly convincing but artificially generated. Examples: fake audio of candidates engaging in corrupt behavior, deepfake docs, and beyond.
- This will be magnified by viral cheap fakes, which include real video that is manipulated with AI to portray something that didn’t happen. Such cheap fakes are already cropping up in the public discourse and manipulating public opinion.
- Tech developments will mean that interactive, deepfake audio will be on the rise by November, e.g. the ability for an attacker to make a phone call purporting to be an election official or candidate that is lifelike, interactive, and indistinguishable from reality. This may include false flags intended to allow one side to justify an overreaction.
- Experts are concerned about targeting of vulnerable populations. For instance, it will be easier for adversaries to target non-native English speakers and evade detection by election authorities.
Participants are collaborating on an edgy, national public awareness campaign. Groups agreed that “rapidly elevating public awareness” was a crucial defense strategy. Namely, voters must be educated — especially in crucial states — about deepfakes designed to confuse them about time, manner, and place of voting and provided easy resources to validate true information.
- Represent.Us is leading the production of an initial PSA spot, with the help of top celebrity influencers, to cleverly educate voters
- An array of civil society groups are planning to join forces to support this public awareness campaign — to be called “Can’tFake.Us” — with the assistance of social media companies.
- For more information, see the draft PSA we produced about the threat, NBC’s coverage of our previous war game, and Axios’ reporting on our work with Hollywood creatives on combating the threat.
|
Sydney Owens 200 Massachusetts Ave NW, Washington DC, Unsubscribe |
|
|
|
