A.I. use in political ads would have to be disclosed under new bill

ALBANY — Political campaign ads, speeches and robocalls to voters created through artificial intelligence would have to be clearly identified as such under state legislation introduced last month.

The measure stems from the explosive use of artificial intelligence to influence consumers and voters.

“If a candidate wants to be seen riding in on a white horse, that’s possible now,” said Assemb. Clyde Vanel (D-Queens), sponsor of the bill. “If they want to be seen as heroic, they could show themselves leaping over a building in a single bound and saving a cat … it’s not evil fake, but we should know it's A.I.”

Aiwei Lyu, an innovation professor for the State University of New York and director of the University at Buffalo Media Forensic Lab, said the proposed legislation would help, “control the unregulated use of generative A.I. and synthetic media in politically oriented ads.”

Vanel’s bill would require disclosure that the technology was used in campaigning, similar to required disclosure of the source of funding for political ads. The bill could be debated in the 2024 legislative session beginning in January.

Under the proposal, political ads or messages created with the assistance of A.I. would need to carry a disclaimer: “This communication was generated using artificial intelligence." Violations could result in civil penalties of up to $1,000 per offense or the cost of producing the communication, whichever is greater.

“We want voters making informed choices based on issues that are important to them,” said the bill’s Senate sponsor, Sen. Jake Ashby (R-Castleton). “It’s on government to make sure that new innovations in technology and communication strengthen our democracy, not undermine it.”

Bipartisan sponsorship on statewide issues is uncommon in Albany and the Republican’s sponsorship may be a hurdle. The bills with the best chance of progressing are those sponsored by Democrats, who control the Senate and Assembly.

“I’m hopeful that my colleagues on both sides of the aisle will agree that New York politics is sufficiently crazy already,” Ashby said in a statement. “We don’t need to throw unregulated A.I. in the mix.”

The bill isn’t specifically aimed at deepfake images that could be used to distort the image or speech of political opponents. Such images would be more directly covered under a bill approved in June and sponsored by Sen. Michelle Hinchey (D-Saugerties) and Assemb. Amy Paulin (D-Scarsdale). The measure awaits Gov. Kathy Hochul’s signature or her veto.

In February, researchers at Stanford University reported on results of research into the ability of A.I. to change voters’ minds on issues such as assault weapon bans and taxes on carbon-emitting devices to help combat climate change.

The research found A.I. writing could influence viewers to change their views on hot-button issues.

“Clearly, A.I. has reached a level of sophistication that raises some high-stakes questions for policy and lawmakers that demand their attention,” said Robb Willer, a professor of sociology, psychology, and organizational behavior who was part of the study.

By Michael Gormley

michael.gormley@newsday.com@GormleyAlbany

Michael Gormley has worked for Newsday since 2013, covering state government, politics and issues. He has covered Albany since 2001.