Do I Have to Disclose AI-Generated Content?

The FTC Rule: Sponsored Content and Meaningful AI Use

The FTC's Endorsement Guides require disclosure when two things are true at the same time: a post is sponsored (through payment, a free product, affiliate commission, or another material benefit), AND AI generated a meaningful part of the content. Spell-check and grammar tools do not count as meaningful AI use. But if AI was used to write a product review, generate images, create a video script, or substantially draft social media copy, AI involvement must be disclosed separately from the sponsorship disclosure.

Both disclosures must be clear and easy to see without clicking. In videos, they must appear on screen in the first few seconds. In text posts, use simple language like 'Sponsored: This post was written with AI help.' The FTC updated its guidance in May 2026 and is actively enforcing these rules. Each non-compliant post can draw a separate penalty, so the stakes are real for all creators and influencers.

If a post is unpaid and reflects only personal opinion, the FTC Endorsement Guides generally do not require disclosure, even if AI was used to write it. The disclosure requirement applies because of the sponsorship or material benefit, not because AI was involved.

EU AI Act Article 50: Starting August 2, 2026

If people in the European Union can see content, Article 50 of the EU AI Act requires disclosure when content is artificially generated or manipulated. This applies to text, images, audio, and video published online. Location of the creator or business does not matter. Sponsorship status does not matter. Account size does not matter. If content is AI-generated and accessible to EU users, Article 50 requires disclosure.

The main exception is AI-generated text meant to inform the public: if a real person substantially reviewed it and took documented editorial responsibility for it, disclosure may not be required. But deepfakes always require disclosure, no matter what.

Article 50 also requires machine-readable marking, meaning the content must be tagged so detection tools can identify it as AI-generated. A text label is not enough by itself; metadata is required. This law takes effect August 2, 2026. Generative AI systems already on the market have until December 2, 2026 to comply.

New York Synthetic Performer Rule: Already in Effect

New York's law (S.8420-A) focuses on a specific use of AI: ads that feature synthetic performers, meaning AI-generated people who look or sound human but do not exist. If ads that reach New York consumers include synthetic performers, a conspicuous disclosure in the ad itself is required. This rule has been in effect since June 9, 2026.

The law carves out some exceptions: audio-only ads, AI used only for translation, and ads for creative works where the synthetic performer use matches what appears in the actual work. But if synthetic performers are used knowingly in ads visible to New York audiences, disclosure is required.

California Bots and the August 2 AI Transparency Deadline

California has two separate AI disclosure rules. The first is the Bot Act (in place since 2019): if a chatbot or AI agent communicates with someone in California with intent to mislead about artificial identity in order to incentivize a purchase, sale, or vote, a clear, conspicuous, in-conversation disclosure that it is a bot is required. The statute provides that a person using a bot is not liable under this section if the person discloses that it is a bot.

The second is SB 942, the AI Transparency Act, which takes effect August 2, 2026. SB 942 targets companies that build large generative AI systems with over one million monthly users. It does not apply to small businesses and creators using those systems. If you use ChatGPT or Claude to write a post, SB 942 does not apply to you. It applies to OpenAI, Anthropic, and similar system providers.

How to Figure Out What Applies to You

Start by asking: Is this post, ad, or conversation sponsored? If yes, disclosure of the sponsorship is required. Then ask: Did AI generate substantive content (not just a spelling fix)? If yes, separate disclosure of the AI involvement is required. This applies to every platform and post.

Next: Can people in the EU see this content? If yes, compliance with Article 50 is required as of August 2, 2026. Disclosure that content is AI-generated and machine-readable metadata are required.

Third: Is this an ad with an AI-generated person in it, and could someone in New York see it? If yes, disclosure of the synthetic performer is required.

Fourth: Is a chatbot being used to sell or influence someone in California, with intent to mislead about the artificial identity? If yes, disclosure that it is a bot, in the conversation itself, is required.

When uncertain whether a particular piece of AI help counts as meaningful, disclosure of the AI involvement is the safer choice. Regulators typically focus enforcement on deliberate non-disclosure rather than overcautious disclosure.

Frequently asked questions