The Maze: Pennsylvania, a US state, has enacted disclosure rules for certain misleading advertisements made with artificial intelligence. Governor Josh Shapiro signed the law on October 22. Advertisers face a December 21 start date, but the rule contains important exceptions: using AI does not automatically make every ad subject to a label.
The Details:
The buyer's impression is the trigger. The Artificial Intelligence Deceptive Advertising Act covers ads using generated or significantly AI-modified content in a way that creates a false impression about the advertised goods or services. Its focus is what buyers believe about a product's nature, origin or authenticity. That matters for retailers turning product descriptions into invented images: the image may look like evidence of a real item. The bill's sponsor, state senator Nick Pisciottano, used that example to explain the consumer-protection goal. It is a rationale for the law, not evidence that labels will reduce scams. The text does not settle every borderline case of what counts as a significant modification or false impression.
The disclosure must travel with the ad. When required, it must identify AI generation or modification from the first time consumers encounter the content. It must use the same medium and be permanent or exceptionally difficult to remove where technically feasible. The law also spells out readable size, contrast and display time, and understandable audio volume and pace. In practical terms, a label has to work in the ad people actually receive. A disclosure tucked away elsewhere cannot simply substitute for that presentation. Advertisers therefore face a delivery question as well as a production question: does the explanation remain noticeable in the format shown to the shopper?
Exceptions and civil enforcement narrow the headline. Text-only content is exempt, even though the definition of synthetic content includes text. Evidently fictional depictions, parody, satire, commentary and criticism are also excluded. The state attorney general or relevant district attorney can seek a court order stopping a prohibited practice and bring a civil case. A court may impose a fine of up to US$3,000, plus legal fees and court costs. That is an authorized ceiling, not an automatic charge for using AI. The law takes effect 60 days after signing. It also preserves other legal recourse, so an AI disclosure does not make a deceptive product claim acceptable.
MarketMaze Take: The commercial change is a connection between creative work and ad delivery. A business cannot assess a covered campaign only by asking which software made its images. It must also consider what the images imply and how the disclosure reaches the buyer. That makes the final advertisement the meaningful unit of review.
A state rule can affect shared creative. Advertisers outside Pennsylvania using the same campaign there have a concrete reason to examine their assets. The selected report notes that labels may appear in other states when a commercial also runs in Pennsylvania. This is a possible consequence of shared advertising, not a nationwide legal requirement or evidence of a global rollout.
The difficult boundary remains judgment. A business might choose broader labelling for simplicity, but the enacted rule has a deception-related trigger and express exemptions. Neither its existence nor a label proves a product image is truthful. The inspected evidence establishes the obligation and its limits; it does not yet establish compliance costs or enforcement outcomes.
