Pennsylvania Enacts AI Ad Disclosure Law With Big Loopholes

Pennsylvania Governor Josh Shapiro signed a new law on Thursday to rein in AI ads. The measure is called the Artificial Intelligence Deceptive Advertising Act (Senate Bill 806). It forces companies to put clear labels on ads made or altered with AI. The rule covers video clips, radio spots, still images, and text.

The law hits multi-million-dollar campaigns, like Coca-Cola’s AI holiday TV commercials. It also covers local diners sharing fake photos of fried chicken on social media. Any ad shown to people in Pennsylvania must follow the rule, no matter where an agency made it.

Pennsylvania joins Utah and California as an early state to police commercial AI. Yet the bill leaves big loopholes. Defense lawyers already see easy ways around the rules.

Technical Factsheet: Pennsylvania SB 806

How the Disclosure Rule Works in Practice

The new law focuses on “synthetic content.” The bill defines this as any media significantly altered or built by artificial intelligence. When an ad uses this content, the sponsor must add a clear label. The label must state that AI built or changed the ad.

The bill sets three clear rules for these labels:

• Format Match: The label must match the media format. A radio spot must say the disclaimer aloud. A video ad must show clear text on screen. A print flyer must print readable words.
• First Sight: The label must show up the moment the AI media appears. Brands cannot bury the notice in end-credits.
• Tough to Strip: The label must stay stuck to the file. It must be hard to crop out or hide during distribution.

The Interstate Broadcast Dilemma

The biggest punch in SB 806 comes from its broad reach. The law applies to any ad shown in Pennsylvania. It does not matter where the ad agency sits.

This rule creates a headache for national marketing teams. A brand might buy an ad slot on a national football game. Or it might run video ads on Hulu and YouTube. Those streams reach homes in Philadelphia, Pittsburgh, and Scranton.

Brands have two choices. They can build complex web tools to show labels only to Pennsylvania IP addresses. Or they can slap the AI label across all national ads. Many teams will choose national labels just to avoid legal traps.

Four Big Loopholes in the Law

Lawmakers want to stop scams. Yet defense lawyers have already found four big gray areas in the text:

1. The “False Perception” Trap

The law does not require labels on every single AI ad. Instead, it sets a strict test. The AI media must create a “false consumer perception about the nature, origin or authenticity” of the product.

This single phrase creates a huge legal loophole. Take a car commercial. An agency might use AI to paint a moody mountain road behind a real truck. Does the fake mountain trick buyers about the truck itself? Or take a soft drink ad that shows talking cartoon animals. Buyers know animals do not drink soda. Lawyers can argue that artistic AI tricks do not fool anyone about the real product.

2. No Clear Line for “Significantly Modified”

The bill covers media that has been “significantly modified or generated by artificial intelligence.” Yet the law gives no clear line for what counts as significant.

Modern photo and video tools use machine learning every day. Photographers use tools in Adobe Photoshop to erase power lines or smooth studio backdrops. Video editors use AI to clean up grain on camera footage. Do these small edits break the law? Or does the law apply only to full AI prompts? Nobody knows yet.

3. The “Evidently Fictional” Pass

SB 806 explicitly exempts parody, satire, commentary, and anything that is “evidently fictional.”

This sounds simple on paper, but generative AI blurs the line between real and fake. Suppose a burger joint uses AI to show a giant cheeseburger. Is that an “evidently fictional” work of art? Or is it a fake photo of food they serve in a box? Courts will have to judge each case one by one.

4. No Private Lawsuits from Consumers

Regular consumers cannot sue companies under SB 806. Only the state Attorney General and local district attorneys have the power to file charges.

State prosecutors work with tight budgets. They must pick their battles. They will likely go after multi-million-dollar scams or massive national brands. Small diners posting fake food photos on Facebook will likely slip right through the cracks.

What Ad-Tech Teams Must Build Next

Engineering teams cannot wait for the first court case. Ad servers, content pipelines, and creative shops must update their software now.

Three technical tools will lead this shift:

• C2PA Content Credentials: C2PA standards embed digital tags into photos and video. Software can read these tags to flag AI media automatically.
• Regional Ad Injection: Ad servers like Google Ad Manager and Meta need rules to check viewer zip codes. They can append required warning banners before serving ads into Pennsylvania.
• Agency Deliverable Audits: Brands must check files from external ad agencies. Teams need written proof of whether vendors used generative tools during shoots.

What Comes Next for Marketers

The Pennsylvania law takes effect 60 days after signing, in late December 2026. Violators face court injunctions and fines up to $3,000 per violation.

The federal government has yet to pass a nationwide AI ad law. In the meantime, individual states are writing their own rulebooks. Brands that run ads in late 2026 must check their media assets today. Adding simple tags now costs far less than fighting state prosecutors later.