AI Disclosure Law Clears New York’s Legislature With Key Wins for Radio, Podcasts.

New York has taken, according to its supporters, a historic step marking a new era of transparency in the use of artificial intelligence in advertising. But in an eleventh-hour move, state lawmakers made a significant change that will impact all audio advertising. They amended the bills (S.1228C/A.606B) to remove audio-only advertising, whether it is over-the-air radio, digital streaming, or podcasts, from the proposal. The update also clarifies that AI-generated language translations are not considered a “synthetic performer.” 

The New York State Broadcasters Association (NYSBA) credits a “series of negotiations and long nights” for the changes. The trade group has been among the critics of the legislation, even launching a media campaign to get listeners and viewers in the effort.

The proposed law seeks to create a “synthetic performer disclosure requirement” that would have required any AI voice be labeled just like any image in TV, print or digital ads so that the average person will be able to quickly decipher the person isn’t real. Failure to properly label a synthetic performer will result in fines of $1,000 for the first violation, and up to $5,000 for subsequent violations.

For television stations and other media outlets that are still subject to the rules, NYSBA says the revised bill also removes any liability for stations, saying the obligation to include a disclosure for using a computer-generated “synthetic performer” rests solely with the creator or producer of the advertisement. Earlier versions potentially make local stations responsible, which NYSBA said would have presented stations with an impossible task of pre-screening and placing disclosures on ads received from other program suppliers. Now, the only time a TV station will be on the hook for the disclosure is if it produces the ad for a client and uses a computer-generated “synthetic performer.”

“While not perfect, the legislation has come a long way,” NYSBA tells members. Among its continued worries is the use of what it says is an “extremely broad” definition as to what constitutes a synthetic performer and it assumes “synthetic performers” are inherently deceptive. “The real issue is the truthfulness of what is being said in an advertisement, not whether the character making the statement is computer-generated or a real human,” the group says in an email.

SAG-AFTRA is celebrating the bill, which it says is the first law in the country, and possibly the world, to mandate disclaimers when consumers are being sold to by AI-generated humans, or “synthetic performers.” National Executive Director Duncan Crabtree-Ireland said that in the age of deepfakes, it’s essential that consumers be informed when an individual they’re seeing in advertisements is not an authentic human. “We must continue to acknowledge and defend the value of real human expertise and artistry in all media and creative work,” she said.

But other groups, including the National Association of Broadcasters, the Association of National Advertisers, and the American Association of Advertising Agencies have criticized the effort. They have said it would have a “substantial” impact on advertising while also injecting “compliance uncertainty” into the process.

The legislation now goes to Gov. Kathy Hochul, who has not said whether she will sign it. If she does, the bill will give media companies 180 days before it becomes effective.

Meantime, a proposal remains pending in Congress that would put a 10-year moratorium on any state-level AI regulations.