Pennsylvania makes early move on AI ad labels as Shapiro signs deceptive advertising law

Pennsylvania Gov. Josh Shapiro signed a new law on Thursday that requires labels on ads generated by artificial intelligence or significantly altered by it, putting the state among the first in the U.S. to demand disclosure for synthetic promotional content.

The Artificial Intelligence Deceptive Advertising Act covers still images, video, audio and text, and it applies to everything from large-company TV spots to small-business flyers. The law matters because it tries to draw a line around AI-made marketing that could mislead consumers about a product or service, even when the ad was created outside Pennsylvania but is expected to run there.

What Pennsylvania changed

The new law requires a disclosure when advertising content has been generated by AI or significantly modified by AI. According to reporting by The Verge, the label must appear in the same format as the ad itself, which means the notice should be visible or audible in the medium where the advertisement appears rather than tucked away in separate fine print.

The statute reaches across formats. That means the same basic requirement can apply to a television commercial, a social clip, an audio spot, an online banner, a printed flyer or a text-based promotion if the material falls within the law’s scope.

It also reaches beyond state borders. Ads produced elsewhere may still need the disclosure if they are expected to air or circulate in Pennsylvania, a detail that gives the law a potentially wider footprint than a simple in-state marketing rule.

Which kinds of ads are covered?

As described by The Verge, the law covers still images, video, audio and text. That broad list is important because AI is now used not only for photorealistic pictures and synthetic voices, but also for scripts, captions, copy and rapid design edits that can change the meaning of a campaign.

• Still images
• Video
• Audio
• Text

The measure is not limited to one part of the advertising industry. It is written to catch both national campaigns and local promotions, including materials made for small businesses.

Why it matters

This is one of the first state laws aimed specifically at AI-content disclosures in advertising, and that makes Pennsylvania a test case for how regulators may police synthetic media in the commercial sphere. The practical question is not whether AI will remain in advertising — it already is — but how much transparency states can require when consumers encounter those ads in the real world.

For readers, the issue is straightforward: an ad that looks or sounds authentic may not be fully authentic. A voice, face, or image can now be generated or altered so convincingly that viewers may not know whether they are seeing a real person, a real product demonstration or a synthetic stand-in. The law is designed to reduce that confusion.

For marketers, the stakes are operational as well as legal. Companies that make ads with AI tools may need to review creative workflows, disclosure placement and media buys more carefully, especially if a campaign could cross into Pennsylvania. The broad reach of the law means compliance may become part of routine ad production rather than a narrow legal afterthought.

For states, Pennsylvania’s move may become a model or a warning. If the rules prove workable, other legislatures could follow. If they prove too vague to enforce, the law may show how difficult it is to regulate synthetic content with broad concepts and limited definitions.

How did Pennsylvania get here?

The background is a fast-moving policy debate over synthetic media and consumer deception. State-level AI disclosure rules in advertising are still an emerging area, but lawmakers have been reacting to a simple reality: AI can generate or materially alter images, audio, video and text at scale, often quickly and cheaply.

That capability creates obvious benefits for advertisers, from faster production to lower costs. But it also raises the risk of misleading claims, especially when an ad’s appearance suggests a real scene, voice or endorsement that does not exist in the way a viewer would assume.

The sponsor of the bill was state Sen. Nick Pisciottano, according to The Verge. Gov. Josh Shapiro signed the Artificial Intelligence Deceptive Advertising Act into law on Thursday, giving the proposal its final political approval and turning it into state policy.

State advertising rules can also have an extra-territorial effect. A campaign produced outside a state may still be subject to that state’s standards if it is distributed there, and the law’s language reflects that reality. In practice, that means a national brand could face different disclosure demands depending on where the ad is seen or heard.

What the law says — and what it does not say

The law’s core standard is targeted: it limits disclosures to synthetic content that creates a false consumer perception about a product or service. That sounds narrow, but the actual application may be broader than it first appears because ads can mislead in subtle ways, not just through outright falsehoods.

Still, several key phrases remain undefined or at least unclear, and those gaps may determine how aggressive the law becomes in practice.

What do the exemptions cover?

The law makes room for parody, satire, commentary, criticism and evidently fictional content. Those carve-outs matter because advertising and political or cultural commentary can overlap in digital media, and lawmakers often try not to suppress obvious humor or creative expression when they regulate deception.

But those exemptions also reveal the challenge of writing a clean rule for synthetic media. A joke, a spoof or a stylized concept ad may be easy to recognize in one context and harder to distinguish in another. That means the final answer may depend on how regulators, courts or industry lawyers interpret the content’s purpose and audience reaction.

The Verge reported that the law language limits disclosures to synthetic content that creates false consumer perception about a product or service. That framing suggests the statute is aimed at deception, not at every use of AI in creative production.

Who will feel the impact?

The immediate answer is advertisers, agencies and platforms that run paid promotions into Pennsylvania. But the practical audience is wider than that, because compliance may affect production teams, legal departments, media buyers and even small business owners who rely on ready-made design tools.

Large companies are obvious targets because their campaigns are more likely to travel across state lines and appear in multiple formats. At the same time, the law explicitly reaches small-business flyers, so local advertisers cannot assume the rules are only for national brands.

That broad application is one reason the measure could matter well beyond the state’s borders. If an out-of-state agency makes a campaign intended to run in Pennsylvania, that agency may now need to think about disclosure obligations earlier in the creative process.

Timeline

The policy moved from proposal to law with a straightforward sequence, though the broader debate around AI ads is still unfolding.

How might enforcement work?

Enforcement is one of the biggest open questions. The law creates a disclosure requirement, but the hardest part of any such rule is deciding when a piece of content crosses the line from ordinary editing into AI-driven modification that must be labeled.

That question matters because the modern ad workflow often blends human and machine input. A campaign may include generative tools for image cleanup, voice synthesis, script drafting or layout changes, yet not every adjustment will obviously qualify as significant modification. That ambiguity could leave regulators with a fact-by-fact judgment call.

Another unresolved issue is who decides whether an ad creates a false consumer perception. That could involve regulators, courts, or a complaint-driven process, depending on how the law is implemented and challenged.

The vague terms identified by The Verge — including significant modification, evidently fictional and false consumer perception — are likely to become the pressure points. They will determine whether the law functions as a strong disclosure mandate or a broad principle that is hard to apply consistently.

What happens next

For now, Pennsylvania has set a legal standard, but the real work begins in interpretation. The most important developments to watch are how officials define the ambiguous terms in the law and whether they issue guidance that gives advertisers a clearer compliance path.

Also worth watching is whether other states copy the Pennsylvania model. The state is moving early, and early movers often influence the shape of later legislation. If lawmakers elsewhere decide that disclosure is the right approach, Pennsylvania’s version could become a reference point for future bills.

Several practical questions remain open:

1. How Pennsylvania regulators will define “significantly modified” content
2. How officials will interpret “evidently fictional”
3. Who will decide whether an ad creates a false consumer perception
4. How enforcement will work in practice
5. Whether other states will adopt similar disclosure rules

Those questions will decide whether the new law becomes a meaningful guardrail or simply one more compliance requirement in a crowded regulatory field. Either way, Pennsylvania has put itself near the front of the states trying to bring AI advertising into clearer view for consumers.

Background: why synthetic ads are a regulatory problem

Synthetic content can include AI-generated images, voice, video and text, and that mix can be persuasive in ways traditional editing never was. A synthetic spokesperson can look real, sound real and speak with a confidence that blurs the line between illustration and representation.

That is the regulatory problem at the heart of the Pennsylvania law. A consumer can be influenced by a commercial without knowing that the person or scene in the ad was manufactured or heavily altered by software. If the result is a false impression about a product or service, the state now says disclosure is required.

In that sense, Pennsylvania is not trying to ban AI in marketing. It is trying to make the use of AI visible when it could affect consumer understanding. That is a narrower objective than a ban, but it is also a more realistic one in a market where AI tools are already embedded in content creation.

Why this may become a national story

Advertising is one of the easiest places for AI policy to move from theory to practice. It is visible, commercial, and often distributed across state lines, which means a state rule can quickly influence national behavior if companies decide it is easier to standardize one compliance practice than to customize every campaign.

If that happens, Pennsylvania’s law could shape how agencies label synthetic content far beyond the Commonwealth. Even if other states do not immediately pass identical laws, national brands may adopt a cautious approach and treat Pennsylvania as a benchmark for disclosure.

At the same time, the law’s gaps may limit its reach. If the definitions remain fuzzy, companies may be able to argue that certain edits are not significant modifications or that a piece of content is evidently fictional. That tension between ambition and ambiguity will define how much real-world impact the statute has.

For now, the state has taken an early step on a problem that is likely to grow rather than shrink. Pennsylvania’s new law says that when AI helps make an ad look or sound like something it is not, consumers should be told.