Today, Public Citizen and Common Cause submitted petitions to state election officials in Alabama, Louisiana, and Wisconsin calling on them to regulate AI (Artificial Intelligence) ‘deepfakes’ in campaign communications.
The petitions rely on a similar rationale to that of an earlier petition submitted by Public Citizen to the Federal Election Commission (FEC), which argued the agency has statutory authority under the law against “fraudulent misrepresentations” to require federal candidates to disclose the use of false and misleading AI-generated content. Alabama, Louisiana and Wisconsin have similar laws on the books that give their elections officials the authority to regulate deepfakes.
AI-generated content — including computer-generated images and fabricated audio statements from real life candidates — , has become so realistic that many voters may not be able to discern the difference between what is real and what it fake,” said Craig Holman, a government ethics advocate with Public Citizen. “The degree to which the technology has improved — and the speed at which it continues to do so — means that 2024 is going to be the first serious deepfake election.”
“AI deepfakes represent a very clear and present danger to our democracy. The opportunity for deceiving and misleading voters has never been so acute as it is today with the vast improvements in fake computer-generated images and voices,” said Ishan Mehta, the director of the media and democracy program at Common Cause. “And ultimately if voters later realize that they have been duped by false and misleading – yet very convincing – campaign ads, they are going to lose even more confidence in the value of elections.”
In the absence of substantial federal action, other states across the country have shouldered the responsibility for regulating AI deepfakes and made rapid progress towards passing legislation. Although only five states currently ban or regulate deepfakes in election ads – California, Minnesota, Michigan, Texas and Washington – similar legislation has been introduced in at least a dozen more states, as documented in Public Citizen’s state AI legislation tracker. Public Citizen has also produced a suggested model law for states to use to regulate deepfakes in campaigns, emphasizing full disclosure of AI-generated content that falsely depicts what candidates are doing or saying with entirely fabricated computer images and voices.
Ilana Beller, a democracy advocate with Public Citizen, added: “If voters believe elections are being decided upon fake news and misinformation, democracy itself is at risk.”
