Three words came out of Pennsylvania’s AI advertising bill on the state Senate floor in June: “knowingly, recklessly or willfully.” The version without them is now law, and companies that put AI-generated or heavily AI-edited images, video or audio in front of Pennsylvania shoppers have until Dec. 21 to work out what it asks of them.
Gov. Josh Shapiro signed the Artificial Intelligence Deceptive Advertising Act, Senate Bill 806, on Thursday. It was one of 58 bills he signed that day and is now Act 99 of 2026. The Senate passed it 48-2 in June. The House approved the same text 163-40 on Oct. 7, and the act takes effect 60 days after signing.
When a label is required
The trigger is narrower than a blanket AI tag. An ad needs a disclosure when it uses synthetic content, meaning material “significantly modified or generated by artificial intelligence,” in a way that creates a false consumer perception about the nature, origin or authenticity of what’s being sold. The bill text also scripts the label. It has to say the ad contains content modified or generated using AI, show up the first time the content is presented, run in the same medium, and be permanent or extraordinarily difficult to remove where that’s technically feasible.
Content that is evidently fictional is exempt, as are parody, satire, commentary and criticism. So is content limited to text, which leaves a text-only ad written with a chatbot outside the law. Adding a label doesn’t protect a misleading ad, either, since the act says disclosure can’t be read to limit any other recourse under law.
Sen. Nick Pisciottano, the sponsor, framed it as consumer protection. “Consumers deserve to know if AI is being used to create a false impression about a product or service,” he said in an Oct. 7 statement. His office’s example was an AI-generated image that could lead buyers to believe they are looking at the actual product. WebProNews reported this week on TikTok accounts promoting GLP-1 apps with AI-generated women and no ad or AI labels.
The three words that got cut
Pisciottano’s own amendment, A03593, struck the knowledge language on June 10. NetChoice, a trade group whose members include platforms, marketplaces and ad services, objected to the change in Oct. 4 testimony to a House committee. “Liability can now attach even where a party had no idea the content was AI-generated, which is close to strict liability,” the group wrote. It asked for the standard back and a one-year runway. The House passed the bill without either change three days later.
New York took a different route. Its synthetic performer law, in effect since June, covers ads featuring AI-generated people, applies only where the advertiser has actual knowledge, exempts audio ads and sets penalties of $1,000 for a first violation and $5,000 after that, under Section 396-b of its General Business Law. Pennsylvania’s law never mentions performers, and it does reach audio. In a client note, Loeb & Loeb said an AI spokesperson giving accurate information may not need a Pennsylvania label, while an AI-altered real person delivering a fake endorsement may. The firm expects many advertisers to label all AI imagery, video and audio in Pennsylvania ads rather than apply the test case by case.
There is no private right of action. Enforcement belongs to the attorney general and the state’s 67 district attorneys, who can seek injunctions or sue. A court can add a civil fine of up to $3,000, plus attorney fees and court costs.