Pennsylvania Governor Josh Shapiro on Thursday signed the Artificial Intelligence Deceptive Advertising Act, a law that requires a disclosure on advertising that has been significantly generated or modified by artificial intelligence, The Verge reported on Saturday. Pennsylvania is one of the first U.S. states to require such labels, and the rule applies to ads of any size, from national television commercials to a restaurant’s flyer. The wording leaves room for interpretation, which is likely to matter once advertisers and regulators start applying it.
What the law requires
According to The Verge, the law covers any ad, whether a still image, video, audio or text, that “has been significantly modified or generated by artificial intelligence.” Such an ad must carry a label in the same format as the ad itself, so a video ad needs an on-screen notice and an audio spot needs a spoken one.
The measure began as Senate Bill 806, sponsored by Sen. Nick Pisciottano, a Democrat from Allegheny County. His office said in May that the bill was reported out of the Senate Communications and Technology Committee unanimously, with bipartisan support, and described the aim as requiring disclosure when AI is used to generate images, audio, video and text tied to the sale of consumer goods and services. The bill passed the legislature on October 7, according to MediaPost.
Formatting rules
MediaPost’s report on the bill, written before the signing, describes how the disclosures must look. Audio disclaimers must be delivered at a “volume and cadence” that consumers can hear, and visual disclaimers must stay on screen long enough to be read. The disclaimer has to be in the same medium as the ad and, where feasible, should be “permanent or extraordinarily difficult to remove from the synthetic content.” We have not read the final signed text, so these details reflect the bill as reported at passage.
Who is covered
The Verge notes that the labels are required regardless of where an ad was created. A commercial produced in another state would still need to carry the disclosure if it is also expected to air in Pennsylvania. For national advertisers, that makes a state rule a practical national concern, much as state privacy laws have shaped nationwide data practices.
Where critics see loopholes
The Verge’s article argues that the law “leaves a lot of loopholes.” It lists several:
- What counts as “significantly modified” is subjective, and the law does not appear to settle it.
- There is a carve-out for “parody, satire, commentary or criticism,” and for anything “evidently fictional,” a phrase whose meaning is unclear.
- The wording suggests a disclosure is required only for content that “uses synthetic content in a manner that creates a false consumer perception” about a product or service. That raises questions about what qualifies as “false” and who decides.
Sen. Pisciottano’s office did not immediately respond to The Verge’s request for comment. In a statement quoted by MediaPost, he said the bill “sets a commonsense transparency standard” so people can make informed choices while still letting businesses use the technology.
Industry opposition
The technology trade group NetChoice opposed the bill. In an October 4 letter to lawmakers, quoted by MediaPost, it argued that federal and state laws already prohibit deceptive advertising and that “layering a parallel, AI-specific regime” on top of them does not improve consumer protection. It also warned that the prescriptive format rules “risk overwhelming short videos, small display ads, and audio spots.”
MediaPost added that Pennsylvania is not alone: last month California Governor Gavin Newsom signed a law requiring ads that use “synthetic” performers to disclose clearly that the digital figures were generated by AI. The state approach is spreading while there is no single federal standard. For a wider comparison, see our global map of AI regulation.
Labels versus provenance
A disclosure label is a legal duty placed on the advertiser. It is different from technical provenance, which attaches verifiable metadata to a file, as described in our explainer on content provenance and C2PA. A label can be removed or ignored by someone who does not follow the law, which is why the Pennsylvania text’s call for disclosures that are hard to remove is notable. For marketers, the practical effect is a new review step, which we discuss in how AI is changing marketing and advertising.
What to watch next
- The official text of the signed act and its effective date, which the coverage we reviewed did not state.
- Who enforces the rule, and what penalties apply, which also were not reported.
- Whether advertisers and trade groups challenge the law or ask for guidance on “significantly modified.”
- Whether other states copy Pennsylvania’s wording or the California approach to synthetic performers.
- How platforms and broadcasters change their ad review for creative made with generative tools.
Key takeaways
- Pennsylvania has signed a law requiring labels on ads significantly generated or modified by AI.
- Labels must match the ad’s format, and the rule applies regardless of where the ad was made.
- Exceptions for parody, satire, commentary, criticism and “evidently fictional” content leave open questions.
- Industry group NetChoice argued existing deception laws already cover the problem.
- California passed a related law last month on synthetic performers, so a state-by-state patchwork is forming.
Frequently asked questions
Does the law ban AI-generated ads?
No. Based on the reports we reviewed, it requires disclosure; it does not prohibit the use of AI in advertising.
Does it apply to ads made outside Pennsylvania?
The Verge reports that labels are required regardless of where the ad was created, so ads that are also expected to air in Pennsylvania may need them.
When does it take effect?
The coverage we reviewed did not give an effective date. Check the text of the act on the Pennsylvania General Assembly website.
