SiteProof AI Blog
AI Content Disclosure Requirements in 2026: What Websites Must Label
Websites must disclose AI-generated content when it could mislead users into thinking it was created by a human. The EU AI Act Article 50, FTC guidelines, and multiple US state laws require clear labeling of AI-generated text, images, and audio in public-facing contexts. Requirements are effective August 2, 2026 under the EU AI Act.
What Counts as AI-Generated Content Under the Law?
✅ Requires disclosure
- •Blog posts generated by AI published as editorial content
- •Product reviews written by AI
- •Customer testimonials generated by AI
- •AI-generated images in advertising
- •Deepfake audio or video content
- •Chatbot responses presented as human responses
◯ Does not require disclosure
- •AI-assisted spell checking or grammar correction
- •AI-powered search ranking
- •Backend AI tools not visible to users
- •AI translation with clear translation labels already present
EU AI Act Article 50 — Content Disclosure Requirements
Article 50 covers two categories:
Direct interaction disclosure (chatbots)
Users must be informed they are interacting with an AI system before or at the start of the interaction.
AI-generated content labeling
AI systems that generate synthetic content — text, images, audio, or video — must ensure outputs are labeled as artificially generated or manipulated, particularly where this could mislead users.
Effective date: August 2, 2026 for enforcement. Maximum fine: €15 million or 3% of global annual turnover.
FTC Requirements for AI Content Disclosure
The US Federal Trade Commission has issued guidance under Section 5 of the FTC Act requiring that AI-generated endorsements, reviews, and advertising be clearly disclosed. FTC enforcement priorities:
US State AI Content Disclosure Laws
California SB 942 (AI Transparency Act) — effective 2026
Requires providers of AI systems that generate text, images, video, or audio to implement disclosure mechanisms for AI-generated content.
Colorado SB24-205 — effective February 1, 2026
Requires disclosure when AI interacts with consumers, extending to content contexts where users may not realize they are engaging with AI output.
Texas HB149 — effective January 1, 2026
Requires disclosure for AI use in health care content and broader business contexts.
How to Implement AI Content Disclosure on Your Website
For chatbots and interactive AI
For AI-generated written content
For AI-generated images
How to Audit Your Website for AI Content Compliance
For an automated version of this audit, our AI Privacy Scanner checks cookie behavior, privacy policy gaps, and third-party AI data flows automatically.
Ready to check your site? Run a free scan — no signup required.
For the GDPR-specific steps, see our GDPR AI compliance checklist.
Also see: chatbot disclosure requirements.
Frequently Asked Questions
Does every piece of AI-assisted content need a disclosure?
No. The EU AI Act distinguishes between AI-generated content (fully machine-produced) and AI-assisted content (human-edited with AI help). Article 50(4) targets 'synthetic audio, image, video or text content' generated by AI for distribution. Purely human content that was edited or spell-checked with AI tools generally does not require disclosure.
What does a compliant AI content disclosure look like?
A compliant disclosure identifies the content as AI-generated and is visible without the user having to search for it. Common formats: a label near the content ('Generated with AI'), an 'AI disclosure' section in the article, or a site-wide policy statement for AI-heavy content sites. The key is that a typical user would notice it.
Do blog posts written with AI assistance need a disclosure?
This depends on the degree of AI involvement and your jurisdiction. Under current EU AI Act guidance, blog posts that are substantially human-authored and edited likely do not require per-post disclosure. However, if AI generated the post with minimal human editing, disclosure is recommended. The FTC's standard is 'clear and conspicuous' — when in doubt, disclose.
What about AI-generated product descriptions?
EU AI Act Article 50(4) applies to AI-generated text 'intended to inform the public on matters of public interest.' Commercial product descriptions are in a grey area. FTC guidance is more broadly applicable to commercial speech and requires disclosure if AI generation would be 'material' to consumers — i.e., if they would want to know.
Is there a standard label I should use?
There is no single mandated label format. The EU AI Act requires machine-readable marking (C2PA metadata or equivalent) for AI-generated images and audio at scale. For text, a simple visible label works. Many websites use 'AI-generated', 'Written with AI assistance', or 'AI content' labels. Consistency matters more than exact wording.
Check Your Website Now — It's Free
Run a free EU AI Act compliance scan. No signup required.
Start Free Scan →