Pennsylvania has enacted a new disclosure requirement for advertisements made with artificial intelligence, placing the state among the early movers on rules aimed at synthetic media in commercial messaging. Governor Josh Shapiro signed the Artificial Intelligence Deceptive Advertising Act on Thursday, according to reporting by The Verge.
The law requires an advertisement that has been significantly modified or generated by AI to carry a label. The requirement applies across formats: still images, video, audio and text. That makes its reach broader than rules focused only on deepfake video or political communications. A television commercial, a social-media graphic or a local business flyer could all fall within its scope if they meet the law’s threshold.
One notable feature is the required form of the disclosure. The label must be presented in the same format as the advertisement itself. A video ad would therefore need a video-appropriate disclosure, while an audio ad would need an audio disclosure. The provision attempts to prevent an advertiser from burying a notice in a format that consumers are unlikely to encounter while experiencing the ad.
A state rule with effects beyond state lines
The act applies to advertising expected to run in Pennsylvania regardless of where it was created. That could make its effects visible outside the state. National campaigns are often produced once and distributed across many markets, so advertisers may decide that adapting a campaign for one state is less efficient than applying a disclosure more broadly.
That possibility does not mean every AI-assisted ad will suddenly look the same nationwide. Advertisers can still produce Pennsylvania-specific versions, and the ultimate impact will depend on enforcement and on how companies interpret the statute. But the law adds another compliance question to the production process at a time when image, video, voice and text-generation tools are becoming routine parts of advertising workflows.
The policy is directed at consumer perception, not at AI as a creative tool in the abstract. Its language suggests that disclosure is tied to synthetic content used in a way that creates a false consumer perception about a product or service. That focus reflects a central problem for regulators: AI can help make harmless marketing material, but it can also create realistic depictions or claims that make a product appear to do something it cannot do.
Definitions will matter
The practical boundary of the law remains uncertain. The statute refers to content that has been “significantly modified or generated” by artificial intelligence, but what counts as significant is subjective. A campaign made entirely from generated imagery is a clearer case than one in which AI was used to remove an object from a photograph, clean up audio or revise a draft of advertising copy.
That ambiguity matters because AI tools are increasingly embedded in ordinary software. An advertiser may use a system that suggests wording, retouches a product photo or alters a background without treating the output as fully synthetic. The more routine those features become, the harder it may be for businesses to distinguish a modification that requires a label from one that does not.
The act also includes exemptions for parody, satire, commentary and criticism, as well as material that is evidently fictional. Those exceptions recognize that not all synthetic media is intended to persuade consumers about a product. Yet they introduce further questions. Whether something is evidently fictional may depend on context, audience and presentation, and the reporting does not establish how Pennsylvania will apply that standard in disputed cases.
Disclosure is not a complete answer
A label can give audiences useful context, but its effectiveness depends on more than the statute’s existence. It must be noticeable, understandable and delivered at a point when a consumer can use the information. The same-format requirement addresses part of that challenge, particularly for audio and video, but it does not resolve questions about label size, placement, duration or wording described in the available reporting.
There is also a difference between disclosing AI involvement and explaining a claim. A consumer may know that an image or voice was generated or altered while still needing to assess whether the underlying statements about price, performance, availability or endorsements are accurate. The new requirement is therefore best understood as an additional transparency measure rather than a substitute for the broader rules that govern deceptive advertising.
For advertisers, the immediate task is likely to be operational. Agencies and in-house marketing teams will need to track where AI enters the creative process, identify campaigns expected to reach Pennsylvania and decide when a disclosure is appropriate. They will also need to account for the law’s unresolved terms in internal review processes.
An early test for synthetic-media policy
Pennsylvania’s law is significant because it extends the disclosure debate into everyday commercial advertising. It treats AI-generated or heavily modified material as a consumer-transparency issue across multiple media formats, not only as a political deepfake problem.
Its limitations are just as important as its ambitions. The key thresholds described in the law leave room for disagreement, and the exemptions will require interpretation. How state authorities, advertisers and consumers understand those provisions will determine whether the act becomes a narrow rule for conspicuous synthetic deception or a more influential model for AI-advertising disclosure.
For now, the clearest result is simple: advertisements that use AI in Pennsylvania may need to tell audiences so. The harder work lies in deciding exactly which advertisements qualify and what a meaningful label looks like in practice.
This article is based on reporting by The Verge. Read the original article.
Originally published on theverge.com
