Pennsylvania has enacted an artificial-intelligence advertising law, but it is narrower—and in some respects more exacting—than a blanket rule that every AI-created image must carry a warning. The General Assembly records Senate Bill 806 as Act 99 of 2026, dated 22 October. Its operative condition is deception: an advertiser must disclose synthetic content when using it in a way that creates a false consumer perception about the nature, origin or authenticity of the goods or services being promoted.[1]
For a synthetic testimonial or a fabricated product demonstration, the distinction can be consequential. The legal question is not only whether software generated the picture or voice. It is whether the advertisement knowingly, recklessly or wilfully uses that content to create the specified false impression. Disclosure is part of the remedy; it is not a general licence to make otherwise misleading claims.[2]
What advertisers must disclose
The statute defines synthetic content broadly to include information such as an image, video, audio clip or text that AI has significantly generated or modified. Section 4 then describes the prohibited conduct and requires a disclosure stating that content in the advertisement was modified or generated using artificial intelligence. That notice must appear the first time the synthetic content is presented, use the same medium and, where technically feasible, be permanent or exceptionally difficult to remove.[2]
'Clear and conspicuous' is given its own detailed definition. The disclosure must have adequate size, contrast, location, duration and audibility for the medium. It should not contradict other information, must be close to the claim it qualifies, and must be communicated in a language or interaction the consumer can understand. These are practical production requirements, not merely a direction to place an unreadable AI label somewhere in the footer.[2]
Act 99: the scope of the disclosure rule
| Issue | What the enacted text says |
|---|---|
| Trigger | Deceptive synthetic advertising creating a false impression about goods or services |
| Disclosure | At first presentation, in the same medium; difficult to remove where feasible |
| Exceptions | Evident fiction; parody, satire, commentary or criticism; text-only content |
| Enforcement | Injunctions and civil actions; court fine up to $3,000 |
| Effective date | 21 December 2026: 60 days after enactment on 22 October |
Three exemptions limit the rule: content evidently fictional from its context; parody, satire, commentary or criticism; and content limited to text. The last matters particularly for copywriting tools. A text-only advertisement is outside this specific disclosure subsection, although the act expressly leaves other legal recourse available. This is not a finding that a false text claim is lawful under consumer-protection law.[2]
How the rule is enforced
The statute does not ask platforms to watermark every generated piece of content. Nor does it establish a general right to have an AI output removed from the internet. Its enforcement provisions permit the attorney general or an appropriate district attorney to pursue injunctions and civil action, with a court able to award a fine of up to $3,000 as well as actual attorney fees and costs. Those details must not be confused with automatic penalties for any failure to use a generator label.[2]
When does it begin?
The act takes effect 60 days after enactment: 21 December 2026, calculated from the recorded 22 October signing. It is therefore enacted but not yet operative on 26 October. The preparation period gives advertisers time to check how a disclosure will survive different placements, formats and reposting.[2]
Sources & further reading
- Senate Bill 806 and Act 99 of 2026 Pennsylvania General Assembly · Accessed 26 Oct 2026
- Artificial Intelligence Deceptive Advertising Act, enacted bill text PN1808 Pennsylvania General Assembly · Accessed 26 Oct 2026