As lawmakers push to regulate AI in advertising, this New York agency is running AI ads
Curly_photo via Getty Images
The state's Office of Addiction Services and Supports is running television ads featuring AI-generated faces without disclosing the technology to viewers.
This article was originally published by New York Focus.
It’s been difficult to miss this fall’s ad campaign from New York’s Office of Addiction Services and Supports (OASAS). The “Turn the Page on Stigma” campaign has blanketed subway cars, billboards, local television networks, and social media feeds across the state with ads reminding viewers not to judge those battling addiction.
The ads feature people holding up books or emerging out of book pages, with captions such as “Know what they say about judging a book by its cover?”
There is indeed more than meets the eye about the people in these ads: Many are not real, but instead reproductions generated by artificial intelligence.
The campaign comes amid growing efforts by state lawmakers to regulate the use of AI in advertising, which they say can mislead viewers and undermine trust. The ads don’t disclose the use of AI. In a statement, OASAS spokesperson Evan Frost said that OASAS partnered with the Binghamton-based agency Idea Kraft to produce the ads, which were made with a blended approach of AI and traditional methods. A spokesperson for Idea Kraft directed questions back to OASAS.
“The next phase of the ‘Turn the Page’ campaign will feature real individuals in recovery, sharing their own stories in their own voices,” Frost wrote to New York Focus.
Another OASAS spokesperson said that AI was only used in video communications, and that “the remainder of the campaign, including the rest of the video elements, billboards, print, and subway signs, was produced entirely through traditional methods.”
But certain images on OASAS’s website also resemble AI-generated content, said Robert Alexander, a psychology professor at the New York Institute of Technology who uses AI to study human vision. (Frost, the spokesperson, acknowledged that an image on the website was “a screenshot from the TV ad” and did use AI.) Alexander said that he wasn’t sure whether the OASAS subway ads, which portray individuals holding books to their heads, were generated with AI. They appeared “to have been at minimum digitally manipulated,” he said.
“If visitors sense that images depicting patients, families, or care settings are synthetic, the result can be a subtle but significant loss of authenticity,” Alexander said, arguing that AI should be used only with “explicit disclosure and sensitivity to the lived realities of those the agency serves. It’s disheartening to see that that might not be happening here.”
OASAS’s television ads feature a variety of different AI-generated subjects alongside one voiceover. Vaibhav Diwanji, a University of Kansas researcher who has studied AI advertising, said this may be a case of “dynamic content optimization,” which allows advertisers to create several versions of an ad to quickly market to different demographics.
The ad campaign comes amid an explosion in AI video production — and a push to regulate it.
In June, the state legislature passed a bill that would require advertisers to disclose the use of AI-generated humans in commercial advertising.
The bill was sponsored by Assemblymember Linda Rosenthal and State Senator Michael Gianaris. Governor Kathy Hochul is currently weighing whether to sign it.
Because OASAS ads fall under the category of a public service announcement, they would not be required to disclose the use of AI under the bill.
But Rosenthal said the OASAS ads felt dishonest to her. “I don’t see why they can’t use real people,” she said.
At least one other agency has used AI in advertising without disclosing it: The New York Department of Health ran an ad this spring which employed AI, a spokesperson confirmed to New York Focus.
Legislators have also proposed mandating the disclosure of AI in political ads. Andrew Cuomo’s recently-ended campaign for New York City mayor attracted attention this year for using AI in advertising; his ads, unlike the state agencies’, did disclose the use of AI.
As AI spreads, Rosenthal said, it becomes increasingly important for the public to understand what is being shown to them.
“People have a right to know if what they’re looking at is real or not,” she said. “That’s why I think we should proceed cautiously. Especially state agencies.”