New York has passed what is believed to be the first law in the nation that will hold media outlets, including radio stations, responsible for airing political advertisements that include voices and images created with artificial intelligence. The law says they will face legal action if they knowingly air AI-generated content without disclosures, by permitting a candidate whose voice appears in the allegedly deceptive ad to sue the outlet as well as ask a court to block the ad’s distribution and recover court costs and legal fees.
“This audio has been manipulated,” is the disclosure that New York will now require radio stations to air. Similar disclosures for television will swap out the word “audio” for “video,” while newspapers would run disclosures that use the word “image.”
One way media outlets would be able to avoid liability is if they disclose the voice or image is not real. For audio outlets like radio, the law says the ad must clearly state at both the start and finish of the ad that AI was used. And if the ad runs longer than two minutes, the disclosure must be interspersed within the audio by at least every two minutes. It says the disclosures must be in a “pitch that can be easily heard by the average listener” and in the same language as the rest of the audio.
There are some exceptions to the law. Bona fide news reports do not need to have the AI disclosures. Neither does media that “constitutes satire or parody” like what a radio morning show might come up with. Broadcasters would also not be liable if they relay something deceptive that they picked up from another source, like a newspaper or website, that made a good faith effort to verify its authenticity.
Broadcasters See Conflict Between Federal And State Law
The New York State Broadcasters Association says its position is that liability should rest with the creator of the political content, not the broadcaster or any entity that distributes the content. They say the new state law is also running up against federal law, under which stations are required to broadcast a political commercial as they receive it. That includes prohibitions on adding a label or disclosure, which NYSBA says sets up a direct conflict between federal and state law with any political advertisements stations receive directly from a federal candidate's campaign committee. And it says the law is overly vague about what constitutes deceptive manipulation, saying any changes in color or selecting parts of a speech could be considered a violation.
NYSBA had been working with lawmakers on the bill since February when it was introduced. But it was suddenly dropped into the state budget last month, and it quickly became law. The broadcast trade group says the carve out for newscasts is “no exemption at all” and that the law puts broadcasters in a tough spot. “Stations have no way of knowing if the content they receive from third parties contains deceptive AI,” NYSBA says in an update to members.
Broadcasters have been told by state lawmakers that the language in the new law may be revised, and that meetings with legislative allies are in the works. In the meantime, NYSBA is making several recommendations. That includes the suggestion that stations obtain a letter in writing from any political content provider, whether advertising or other content, that says what they are providing the station does not contain any materially deceptive AI-generated material. It is also drafting a form that stations can use. And NYSBA tells members that if they are unable to obtain a letter from the ad buyer, they should opt to run the AI disclosure in any political content.
The Albany law firm Greenberg Traurig says the issue came to the forefront after an audio clip circulated in New York City earlier this year which appeared to be the voice of a political party chairman using profanities to describe an incumbent member of Congress. But the viral clip turned out to be AI-generated, which attorney say highlights the way AI technology may be used deceptively to impact politics and elections.
“New York’s recent legislation may indicate a trend of AI regulation across the country,” it says in a blog post. “As AI technology continues to improve, it has become increasingly difficult to determine whether media is real or digitally altered. This has resulted in the spread of misinformation and interference with elections.”
NYSBA will hold a webinar for New York broadcasters to discuss the new law. Broadcasters can register HERE.