Pennsylvania Now Requires Labels on AI-Generated Ads

What happened

Pennsylvania Governor Josh Shapiro has signed the Artificial Intelligence Deceptive Advertising Act into law. According to The Verge, the law requires any ad that "has been significantly modified or generated by artificial intelligence" to carry a label. That covers still images, video, audio and text.

The label has to appear in the same format as the ad. A video ad needs a disclosure inside the video. An audio spot needs one in the audio. The law applies to a national TV commercial and equally to a restaurant's crowded flyer promoting wing specials.

Pennsylvania is one of the first states to require this kind of disclosure. The requirement also applies regardless of where the ad was created. If a commercial made elsewhere is expected to air in Pennsylvania, you may see the disclosure appear there too.

The loopholes are big

The Verge points out that the law leaves a lot of room for interpretation.

• "Significantly modified" is subjective. Is AI retouching a product photo significant? What about an AI-written headline that a human edited?
• There are carve-outs. Parody, satire, commentary or criticism are exempt, as is anything "evidently fictional." The meaning of that last phrase is unclear.
• False perception may be the trigger. The wording suggests a disclosure is only needed when synthetic content creates a false consumer perception about a product or service. What counts as false, and who decides, is not spelled out.

The bill's sponsor, State Senator Nick Pisciottano, did not immediately respond to The Verge's request for comment, so we have no official clarification yet.

Why it matters for marketers

Even with loopholes, the direction is clear. Regulators are moving from general worries about AI content to specific labeling rules for advertising. Pennsylvania may be early, but other states will be watching.

The cross-state reach matters most for ecommerce and agencies. If you run paid social, video or broadcast campaigns, you rarely control exactly where an ad lands. A campaign built for a national audience can end up in front of Pennsylvania consumers, so a single compliance approach is often easier than state-by-state versions.

There is also a trust angle. Labels on ads could change how audiences react to creative. Some shoppers may not care. Others may read an AI label as a sign of cheap or inauthentic content, especially for products where authenticity sells, like food, beauty, fashion and travel.

What to do now

1. Audit your AI use in ads. List where AI touches your creative: generated images, edited product photos, synthetic voiceovers, AI-written copy, avatar presenters. You cannot decide what needs a label if you do not know what is AI-made.

2. Set an internal definition of "significant." The law does not define it, so write your own rule. For example, treat fully generated visuals or voices as labeled by default, and treat minor cleanup as lower risk. Document the reasoning.

3. Focus on false impressions. Based on the wording reported, the real risk is content that misleads people about what they are buying. An AI-generated product image that does not match the real item is a bigger problem than AI-assisted brainstorming of a slogan.

4. Build label-ready templates. Since the disclosure must match the ad format, plan space for it in video end cards, audio scripts, banner layouts and printed material. Adding it at the last minute usually looks clumsy.

5. Brief your agencies and freelancers. Ask them to flag AI use in every deliverable. Contracts should require disclosure of AI tools used on your creative.

6. Watch for guidance. Because key terms are vague, expect questions and possibly clarifications. Talk to legal counsel before relying on any exemption such as "evidently fictional."

The bottom line

This is not a ban on AI in advertising. It is a transparency rule with fuzzy edges. Businesses that track their AI use now will be better placed when stricter or clearer rules arrive elsewhere.