If you run ads that reach Pennsylvania consumers and any part of the creative came out of an AI tool, you have until December 21 to put a label on it. The state’s new Artificial Intelligence Deceptive Advertising Act does not ask whether you knew the asset was synthetic. It asks whether a buyer could be misled about what they are getting.
“On Thursday, Pennsylvania Governor Josh Shapiro signed the Artificial Intelligence Deceptive Advertising Act into law, requiring disclosures on AI-generated ads. The law requires any ad — whether it’s a still image, video, audio, or text — that ‘has been significantly modified or generated by artificial intelligence’ to include a label in the same format as the ad itself.”
The Verge — Terrence O’Brien, Pennsylvania will require labels on AI-generated ads
Our take
The detail that should move your calendar is not the fine. It is the word that got deleted.
The bill is Senate Bill 806, now Act 99 of 2026. A June amendment struck the phrase “knowingly, recklessly or willfully” from the liability language, and the Senate passed it 48-2 before the House approved the same text 163-40 on October 7 (WebProNews). Strip that phrase out and intent stops being a defense. The trade group NetChoice put the consequence plainly in its objection: “Liability can now attach even where a party had no idea the content was AI-generated.”
That matters because almost nobody running small-business ads today knows, with certainty, which of their assets are synthetic. AI generation is no longer a tool you choose to open. It is baked into the ad platforms themselves, into stock libraries, into the upscaler your designer uses to rescue a low-resolution product photo, into the background-removal button, into the “enhance” slider. A restaurant owner who asked a freelancer for a wing-special flyer has no idea whether the steam coming off those wings was rendered or photographed. Under Act 99, not knowing is no longer the end of the conversation.
The second detail worth reading twice is what the label has to be. It must appear the first time the content is shown, run in the same medium as the content, and be permanent or extraordinarily difficult to remove where that is technically feasible. That rules out most of what marketers would instinctively reach for. A disclosure typed into the caption field of a scheduling tool is not in the same medium as a video. A line added in the comments is not permanent. The label belongs burned into the asset, which means it belongs in the design file, not in the publishing workflow. If you are producing creative in Canva or Adobe and shipping it to five placements, the label has to be in the export, not in the five captions.
Enforcement sits with the state attorney general and Pennsylvania’s 67 district attorneys. A court can order an injunction and impose a civil fine of up to $3,000 plus attorney fees and court costs, and there is no private right of action. Read that as good news and bad news. The good news is you are not facing a plaintiffs’ bar hunting unlabeled flyers. The bad news is that $3,000 is not the real exposure; the injunction and the fee award are. An order to pull a campaign in the third week of December costs a retailer far more than the fine, and the fee-shifting means a dispute you would normally absorb now comes with the other side’s legal bill attached.
One more thing the statute makes explicit: the label does not launder the ad. Disclosure cannot limit other legal recourse, so a misleading ad that carries an AI label is still a misleading ad. Sponsor Sen. Nick Pisciottano framed the intent as narrow — “Consumers deserve to know if AI is being used to create a false impression about a product or service” — and that is the test to keep in your head. The label is a transparency requirement layered on top of deception law you were already subject to, not a safe harbor underneath it.
Where this gets interesting for agencies is contractual. If intent is off the table, the only remaining lever is provenance: who made the asset, with what, and who said so in writing. That is a paperwork problem, and paperwork problems are solvable. Build the record once and the compliance question answers itself every time, in Pennsylvania and in the next state that copies this template — which is the realistic assumption, because Pennsylvania is one of the first states to require disclosures on AI content, not the last.
Where the statute’s edges fall — what counts as “significantly modified,” how the carve-outs for parody, satire, commentary and criticism apply to your category — is a question for your lawyer, not your marketer. Ask it before December 21, not after a demand letter.
What this means for your business
- Inventory every live and scheduled ad before December 15. Pull your active creative out of Google Ads, Meta Ads Manager, your scheduler, and your print vendor’s queue. For each asset, record one field: human-made, AI-generated, or AI-modified. Anything you cannot answer, treat as AI-modified.
- Put the disclosure in the export, not the caption. Add an AI-disclosure layer to your master templates in Canva, Adobe, or whatever produces your finals, so the label ships with the file to every placement. One template change beats labeling every post by hand.
- Turn off, or document, platform auto-generation. Automated creative features that generate or restyle assets on your behalf are the likeliest source of unlabeled synthetic content. Either disable them for Pennsylvania-facing campaigns or confirm in writing what they produce.
- Add an AI-provenance clause to every creative contract. Freelancers, designers, and agencies should represent in writing which deliverables involved AI and indemnify you if that representation is wrong. Do this in your next contract cycle, not after an inquiry.
- Build the process as if it is national. Assume more states adopt a version of this. A provenance field in your asset library and a disclosure layer in your templates cost you one afternoon now and nothing later.
Terrence O’Brien’s full report on the signing is worth reading in its entirety at The Verge.
Put AI to Work for Your Business
MCNM Marketing — the team behind this publication — helps businesses across Las Vegas, Southern California, and Northern Arizona turn stories like this into revenue with AI marketing solutions, SEO & digital strategy, and marketing automation. A+ BBB rated.
Discover more from Digital Media Marketing, and Technology News
Subscribe to get the latest posts sent to your email.

