Pennsylvania: Mandatory Labeling for AI-Generated Advertisements

Pennsylvania has enacted a law requiring a warning on any advertisement generated or significantly modified by artificial intelligence, with a mention in the format of the medium. The provision covers images, videos, audio, and text, and applies even to creations from other states if they target the local audience. Several exceptions and vaguely defined concepts leave room for multiple key interpretations.

Exceptions and Vague Definitions

The law excludes from the disclosure requirement parody, satire, commentary, or criticism, as well as content deemed "clearly fictional." The exact scope of this latter category is not clearly defined. The triggering of the label relies on a "significant modification," a concept presented as subjective. The wording of the text suggests that only the use of synthetic content creating a false perception of a product or service would require disclosure, raising questions about the definition of "false" and the authority responsible for making that determination. The office of Senator Nick Pisciottano, who originated the bill, did not immediately respond to a request for comment.

Obligations for Advertisers Across All Mediums

Any advertisement generated or significantly modified by artificial intelligence must include a label. This mention must adopt the same format as the advertisement medium itself. The obligation applies to still images, videos, audio content, and texts, and targets both television spots from major corporations and promotional flyers from restaurants.

Geographic Scope and Pennsylvania's Position

Labeling is required regardless of where the advertisement is produced. Mentions could thus appear on creations designed outside of Pennsylvania as long as they are intended for distribution there. The state is among the first to impose this type of disclosure. The measure was signed on Thursday by Governor Josh Shapiro.