Pennsylvania signs law requiring labels on AI-generated adverts
Pennsylvania's new law targets adverts that use synthetic content to create a false impression, but leaves key terms open to interpretation.
Pennsylvania has signed a new law requiring many adverts that use synthetic content to carry a clear, conspicuous disclosure. The Artificial Intelligence Deceptive Advertising Act was signed by Governor Josh Shapiro on Thursday, according to reporting by The Verge and the bill text published by Pennsylvania's General Assembly.12
The law covers advertisements that use an image, video, audio clip or text that has been significantly modified or generated by artificial intelligence. Where that synthetic content creates a false consumer perception about the nature, origin or authenticity of the goods or services being advertised, the advertiser must disclose that AI was used.2
That disclosure must appear the first time the synthetic content is shown, use the same medium as the advert and, where technically feasible, be permanent or extraordinarily difficult to remove. A visual notice must be large, clear and visible for long enough to be understood; an audio notice must be delivered at a volume and pace consumers can follow.2
The law takes effect in 60 days. It also gives the Pennsylvania attorney general and district attorneys powers to seek injunctions and bring civil actions. A court may award a civil fine of up to $3,000, alongside actual legal fees and court costs.1
There are important limits. The act excludes content that is evidently fictional, as well as parody, satire, commentary and criticism. It also excludes material limited to text. The phrase “significantly modified” is not defined with a simple technical threshold, leaving advertisers to judge whether an edit crosses the line.12
The rule applies regardless of where an advert was created, so campaigns intended to run in Pennsylvania could affect production and compliance decisions made elsewhere. That does not mean every AI-assisted advert will need a label: the law focuses on synthetic content used in a way that creates a false consumer perception, rather than treating every automated edit as deceptive.12
The practical effect is less a ban on generative tools than a demand for provenance at the point where advertising makes a claim. Brands will need to keep clearer records of how campaign assets were made, decide when an alteration becomes significant and ensure any notice survives the journey from production to publication. The ambiguities leave room for future guidance, enforcement decisions and court tests.
Our opinion
Pennsylvania's approach is sensible in principle because it targets deception rather than creativity. The weakness is its reliance on phrases such as “significantly modified” and “evidently fictional”, which may feel obvious in a courtroom but are much less comfortable for an agency approving a campaign on a deadline. The next useful step would be practical guidance that gives advertisers examples without turning the law into a checklist that can be gamed. For now, the safest approach is to keep an audit trail and disclose synthetic work whenever a reasonable viewer could mistake it for an authentic representation.
Photo by Mikhail Nilov on Pexels.