News

AI is a concern in politics — can Congress solve it?

by Elizabeth Crisp - 09/27/23 12:52 PM ET

NYU Professor Emeritus Gary Marcus, center, speaks before a Senate Judiciary Subcommittee on Privacy, Technology and the Law hearing on artificial intelligence, Tuesday, May 16, 2023, on Capitol Hill in Washington. Seated alongside Marcus are OpenAI CEO Sam Altman, left, and IBM Chief Privacy and Trust Officer Christina Montgomery. (AP Photo/Patrick Semansky)

Imagine it’s two days before voters head to the polls, and an explosive video begins to spread online apparently showing one of the candidates on a racial slur-laden outburst. It sounds and looks just like the candidate.

Only it isn’t real. The whole thing has been created through artificial intelligence (AI) as a “deep fake” to mislead the public.

It could easily tip the scales in a tight election.

In another scenario, imagine getting a robocall from a high-profile politician sharing polling place details. Only they it’s the wrong location and the politician’s voice has been faked.
Those are just a couple of the scenarios that could have a massive impact on the 2024 election cycle as the growth of AI goes largely unchecked.

A hearing will be held Wednesday to examine what effect AI can have on the 2024 election and elections in general.

“I think there’s definitely a realistic path forward,” Sen. Michael Bennet (D-Colo.) told The Hill.

Even with Congress at a stalemate on many issues, the concerns about artificial intelligence have dominated several recent hearings. In Wednesday’s hearing, members of the Senate Rules and Administration Committee will hear mostly from regulators on a proposal to require watermarks or other signifiers when an image or voice isn’t real.

AI has been a topic that has confounded many, much less an aging Congress where members average 64 years old.

“I think the American people are demanding that we do something,” Bennet said.

The hearing will focus on ways to require watermarking and other ways to tip viewers off that they are seeing something that isn’t real. But Bennet and others think a new federal regulatory agency would help keep up with the emerging technology that is creating questions.

Removing regulation from Congress, where bills can take months or years to pass would be a better way to keep up with it, he argues.

Lisa Gilbert, a spokeswoman for Public Citizen, a nonpartisan nonprofit that has been tracking ethics issues in government, said the new tech is coming “so quickly.”

“It could cause huge problems,” she told The Hill.

There has already been an issue with disinformation and skepticism with internet channels, but the AI issue could be different, Gilbert said.

“The problem is already here. The new problem is the believability,” she said.

Watch the hearing here starting at 3:30 p.m. EDT.