Pennsylvania law requires AI ad disclosures starting December

Consumer protection laws often rely on broad language to capture new technologies before they fully develop. In Pennsylvania, a new statute attempts to apply that standard to the rapid rise of artificial intelligence in commercial advertising.

The Artificial Intelligence Deceptive Advertising Act takes effect on December 21. Governor Josh Shapiro signed the bill on October 22. The law requires advertisers to disclose when artificial intelligence technology is used in advertisements within the Commonwealth.

The requirement is not universal. The statute applies specifically to AI-generated content that creates a false consumer perception. It does not cover every minor digital edit or all instances of AI use. The goal is to ensure transparency when the technology might mislead a buyer about the product or service being sold.

Senator Nick Pisciottano, a Democrat representing parts of Allegheny County, sponsored the bill. He argued that distinguishing between real photos and AI-generated images has become increasingly difficult. He stated that the law aims to prevent consumers from being scammed or ripped off by unclear media.

Shapiro described the measure in a Facebook post as a way to give Pennsylvanians clarity. He said the new law ensures that any advertisement using AI technology must make that fact crystal clear. Pisciottano similarly told reporters that a disclaimer must appear if an ad on the internet or social media was generated by AI.

Legal experts note a gap between the sponsors' descriptions and the actual text of the law. Libby O'Neill, an advertising and AI attorney at Loeb & Loeb, pointed out that the statute only requires a label if the AI use creates a false perception. She said the term false perception remains undefined in the legal text.

The ambiguity stems from the bill's drafting history. An earlier version mandated disclosure for all AI content. The final version limited the scope to advertisements and added the false consumer perception standard. O'Neill questioned what work the AI label performs if the ad is otherwise misleading under existing fraud laws.

Eleanor Huyett, an assistant professor at Temple University's Beasley School of Law, advised businesses to err on the side of disclosure. She said that if an AI-generated image makes a product look better than it is in real life, it may trigger the requirement. She suggested that until courts test the law, the conservative path is to label any content significantly modified or generated by AI.

The law leaves open questions about who qualifies as an advertiser. The text does not define the term. Pisciottano argued that the advertiser is the entity creating the ad, such as a tech company using AI tools. He suggested that if a small business uses ChatGPT to make an ad, the tech company would bear the responsibility for disclosure.

O'Neill offered a different view based on the plain text. She said the law applies to advertisers without defining them, which generally includes the business running the ad. She noted that the state attorney general, district attorneys, and courts will ultimately decide who is covered and how the law is enforced.

Violators face civil fines of up to $3,000, plus attorney fees and court costs. The law does not specify if the fine is per violation or per day. The statute places the burden of proof on the enforcement agencies to determine if a false perception was created and whether the responsible party failed to disclose the AI use.