Pennsylvania mandates labeling of AI-generated advertisements
Pennsylvania has enacted a law requiring a label on advertisements modified or generated by artificial intelligence. The text is broadly framed but also leaves several gray areas about what must actually be disclosed.
Pennsylvania now requires advertisements generated or significantly modified by artificial intelligence to display a visible label, following the signing on Thursday of the Artificial Intelligence Deceptive Advertising Act by Governor Josh Shapiro. The law targets ads in image, video, audio, or text form, and applies to both major brands and small businesses. Its purpose is simple: to make synthetic content more transparent in an advertising market where the line between real and fabricated sometimes becomes blurred.
A mandatory label, regardless of format
The law requires that an advertisement that has been "significantly modified or generated by artificial intelligence" display a notice in the same format as the ad itself. In other words, a TV spot, a static visual, an audio file, or sponsored text must carry the same alert if AI played a sufficient role in their production. The text is not limited to content created in Pennsylvania: it also targets ads made elsewhere but intended for distribution in the state.
This point may have impact beyond local borders. A campaign designed for multiple U.S. markets might have to include a compliant version for Pennsylvania, under penalty of appearing there without a label. Marketing departments already dislike small footnotes; here, they rise to the screen.
Why the law still leaves gray areas
The text does not close all doors. First, the notion of "significantly modified" content remains subjective, which may complicate practical application. Second, the law provides exceptions for "parody, satire, commentary or criticism," as well as for any content "evidently fictional." It remains to be seen who decides, and on what basis, when an ad falls into one category or the other.
Another limitation: the disclosure obligation seems to target content that uses "synthetic content" in a way that creates a "false consumer perception" about a product or service. This wording raises several questions [to be verified], notably about the definition of what is "false" and the authority responsible for interpreting it. In practice, the text better protects the principle of transparency than it lays out a perfectly clear enforcement rule.
A signal for other states
Pennsylvania is among the first U.S. states to require disclosure on AI-generated content. The message sent is mainly political and regulatory: AI-assisted advertising will not be left in a completely gray area, even if the contours of control remain imperfect. According to the source, the office of State Senator Nick Pisciottano, author of the bill, had not immediately responded to requests for comment.
For advertisers, the issue is not only legal. It also concerns public trust, in a context where generative tools allow fast, low-cost, and sometimes too fast production. The message from Harrisburg is clear: if an ad has been edited by a machine, it’s best to say so. This is not the end of creative marketing, but it complicates a bit the magic of “we’ll see.”