R.I. legislation targets deepfakes in elections
PROVIDENCE — The peril posed by deepfakes in elections became evident in January when a fake Joe Biden robocall urged people not to vote in New Hampshire’s first-in-the-nation primary.
And while Rhode Island has not seen such a high-profile example, state lawmakers are sponsoring legislation that would regulate deepfakes, or “synthetic media,” in election communications. Deepfakes are fabricated videos, images, or audio created with the use of artificial intelligence that depict a person saying or doing things they never actually said or did.
Representative Jacquelyn Baginski, chairwoman of the state House Innovation, Internet and Technology Committee, and Senate Finance Committee Chairman Louis P. DiPalma have introduced companion bills that would prohibit using “synthetic media” within 90 days of any election unless that media discloses that an image has been manipulated or generated with artificial intelligence.
“The advent and popularity of deepfakes and similar content pose challenges to the functioning of elections and democracy because such communications can deprive the public of the accurate information it needs to make informed decisions in elections,” Rhode Island Secretary of State Gregg M. Amore wrote in a letter to the Senate Judiciary Committee.
But the American Civil Liberties Union of Rhode Island is warning the legislation could impinge on First Amendment rights.
“The bill’s focus on ‘media in election communications’ seeks to regulate speech in the sphere that the First Amendment most fundamentally applies to — the political process,” ACLU of Rhode Island executive director Steven Brown wrote. “In order to ensure that debate on public issues is, in the words of the US Supreme Court, ‘uninhibited, robust, and wide-open,’ the First Amendment provides special protection to even allegedly false statements about public officials and public figures.”
Under the legislation, a candidate whose appearance, actions, or speech are depicted through the use of synthetic media may seek an injunction prohibiting the distribution of the audio or video. The law would not apply to radio and TV stations, websites, and newspapers that report on deepfakes.
Public Citizen, a consumer advocacy think tank based in Washington, D.C., is backing the Rhode Island legislation. It noted Texas, Minnesota, California, Washington, Michigan, and New Mexico already have passed laws regulating the use of deepfakes in elections.
“The rapid advances in deepfake technology have deeply concerning implications for the upcoming elections,” Public Citizens’ Aquene Freechild wrote. “The 2024 election cycle is being called ‘the first AI election” in the United States.”
John M. Marion, executive director of Common Cause Rhode Island, said the federal government has largely ceded regulating such technology so states are stepping up to address deepfakes. He said he is not aware of any deepfakes used in Rhode Island elections, but he said AI technology has come a long way in the past 12 months and he wouldn’t be surprised to see deepfakes in the state’s fall elections.
“Photoshop has been around for years and we expect people might clean up their appearance,” Marion said. But the Rhode Island bill targets election communication that has been manipulated using artificial intelligence and the creation of new content, he said, noting that technology companies have rolled out tools to create fully synthetic video based on simple prompts.
Marion noted that in November, a University of Chicago poll found 58 percent of Americans believe AI will lead to an increase in the spread of misinformation in the upcoming election, and 66 percent support a federal ban on deepfakes.
But the ACLU argued that public officials could easily use the proposed legislation to deter the exercise of free speech. For example, it said political ads often string together comments made at different times by a politician, but if AI is used to create such ads, those ads could face legal action under this bill.
“We don’t wish to minimize the concerns that have generated this legislation,” Brown said, “but we believe that before rushing to regulate this technology in the political sphere, much greater consideration of the ramifications of doing so is needed in order to avoid infringing upon fundamental First Amendment principles.”
Last month, Rhode Island Governor Daniel J. McKee signed an executive order creating the state’s first Artificial Intelligence Task Force and a Center of Excellence for AI and Data within state government. The task force will be chaired by former US representative James R. Langevin, a Rhode Island Democratic who is leading RIC’s new Institute for Cybersecurity & Emerging Technologies.