Pennsylvania Gov. Josh Shapiro signed the Artificial Intelligence Deceptive Advertising Act on October 22, establishing disclosure requirements for certain advertisements using AI-generated or significantly modified content. The requirements take effect December 21, 2026, but do not amount to a blanket labeling mandate for every advertisement made with AI. (Pennsylvania General Assembly)
The central condition is whether synthetic content creates a false consumer perception about the nature, origin or authenticity of the advertised goods or services. The law also expressly exempts text-only content—a significant distinction for businesses using generative AI to write advertising copy rather than produce images, video or audio. (Pennsylvania General Assembly: final bill text)
What triggers disclosure
The General Assembly’s official history confirms approval of Senate Bill 806 on October 22 and identifies printer’s number 1808 as its final text. That record corroborates the signing reported by news outlets and connects the legislative text linked in the initial coverage to the measure that completed the legislative process. (Pennsylvania General Assembly)
Under that text, advertisers cannot create and distribute, or cause the creation and distribution of, covered advertisements without clear and conspicuous disclosure. The obligation concerns synthetic content used to create the specified false perception, rather than the mere presence of an AI tool somewhere in the advertising production process. (Pennsylvania General Assembly: final bill text)
Covered synthetic content includes material generated or significantly modified by AI. Alongside the text-only exception, the measure excludes evidently fictional content and content constituting parody, satire, commentary or criticism. These exceptions matter when assessing whether a particular creative asset falls within the disclosure requirement. (Pennsylvania General Assembly: final bill text)
The Verge’s report broadly includes text among covered formats. However, the final legislative text contains an explicit exception for content limited to text, also identified in independent legal analyses. Advertisers should therefore distinguish the broad definition of synthetic content from the narrower disclosure obligation and its exceptions. (The Verge)
What the notice must communicate
For covered advertisements, the disclosure must state that the advertisement “contains content that was modified or generated using artificial intelligence.” It must be clear and conspicuous, appear when the synthetic content is first presented, and use the same medium as that content. Where technically feasible, it must be permanent or extraordinarily difficult to remove. (Pennsylvania General Assembly: final bill text)
Those requirements address both the message and its presentation. Acknowledging AI involvement is only one part of the obligation: the disclosure also has to accompany the covered content in the prescribed way. The technical-feasibility qualification applies to permanence or difficulty of removal, not to the underlying requirement to disclose. (Pennsylvania General Assembly: final bill text)
Timing and enforcement
The measure provides a 60-day commencement period, placing its effective date on December 21, 2026. Enforcement rests with the Pennsylvania attorney general and district attorneys, who may seek injunctions or bring civil actions. Courts may impose a civil fine of up to $3,000, together with attorney fees and court costs. (Pennsylvania General Assembly)
The act does not create a private right of action. That does not mean individual consumers lose remedies available under other consumer-protection laws: KingSpry business law attorney Karley J. Biggs makes that distinction in an October 20 analysis published before the signing. Her explanation separates enforcement of this particular act from other potential legal claims. (KingSpry)
The unresolved question is scope
In an October 23 analysis, Loeb & Loeb advertising lawyer Libby O’Neill identifies competing readings of the false-perception condition. One would require disclosure when AI depicts a product; another would require a further misrepresentation beyond the fact that AI produced the depiction. These are interpretations of the threshold, not two separate requirements established by the legislation. (Loeb & Loeb)
Biggs likewise advises businesses to assess whether their AI-generated advertising could create a false perception about a product. Her analysis suggests that, pending clarification, businesses may choose a cautious disclosure approach. That is legal commentary about managing uncertainty—not an additional statutory command to label every AI-assisted advertisement. (KingSpry)
Industry guidance also rejects blanket labeling
The development comes as advertising organizations work on their own transparency standards. On January 15, 2026, the Interactive Advertising Bureau released an AI disclosure framework that favors a risk-based approach over universal labeling. Its focus is AI use that materially affects authenticity, identity or representation in ways that could mislead consumers. (Interactive Advertising Bureau)
The framework combines consumer-facing notices with machine-readable information about content provenance. It is industry guidance, not Pennsylvania law. The comparison illustrates an important distinction for marketers and technology builders: a system designed to communicate how an asset was made serves a different purpose from deciding whether a particular advertisement triggers a legal disclosure obligation. (Interactive Advertising Bureau)
For businesses, the immediate significance is therefore narrower than an across-the-board AI labeling rule. The December start date is established, and the notice requirements are specified. Applying the false-perception threshold to particular creative work remains the interpretive issue highlighted by the legal analyses. (Pennsylvania General Assembly: final bill text)
