How to Disclose AI-Generated Content (and Why You Probably Should)
Disclosure is required in some cases, advisable in more, and becoming legally mandated across multiple jurisdictions. Here is what needs a label, what the format should be, and why doing it proactively is better than waiting.
Andy Woods 
AI content disclosure has moved from a niche debate to a legal and platform requirement in 2026. YouTube requires it for certain types of AI content. The EU AI Act mandates it for AI-generated material that could mislead people. TikTok has its own label system. The rules are not uniform, but the direction is clear: disclosure is becoming mandatory in more contexts and the window for treating it as optional is closing.
Here is what you need to disclose, how to do it correctly on each platform, and the argument for doing it even when you are not strictly required to.
What Triggers a Disclosure Requirement
Not all AI involvement in video production requires disclosure. The line is around realistic content that could be mistaken for something it is not.
Requires disclosure:
- Synthetic or AI-cloned voices that sound like real, identifiable people (see also: voice cloning legality in 2026)
- Photorealistic AI-generated footage that depicts real events or real locations
- AI-generated depictions of real people doing or saying things they did not do or say
- AI-generated news or documentary-style footage
Does not require disclosure:
- AI-assisted scriptwriting or editing
- AI-generated background music
- AI narration that is clearly synthetic and not impersonating anyone
- AI tools used for editing, color correction, or post-production
The test is whether a reasonable viewer could be misled about the authenticity of what they are watching. A clearly synthetic narrator voice does not pass that test. A realistic AI-generated clip of a real politician making a statement does.
YouTube's Disclosure System
Disclosure is separate from monetisation. YouTube's inauthentic content policy decides whether AI-assisted videos can earn money at all; the faceless channel guide explains what it covers.
YouTube has a disclosure toggle in YouTube Studio's upload and edit flows. When you upload a video, you are asked whether it contains "realistic altered or synthetic content." If it does, you must check the box.
The label appears as a small note below the video player or in the expanded description. It does not affect views or monetisation directly, though YouTube has indicated that realistic AI content without proper disclosure may be removed.
From May 2026, YouTube's automated detection systems also apply labels to videos it identifies as containing undisclosed AI-generated realistic content. Being auto-labeled is not itself a penalty, but repeated failure to self-disclose when required puts the channel under review.
TikTok's AI Label
TikTok has a separate system. When uploading, there is an option to mark content as "AI-generated." The label appears as a small tag on the video. TikTok has also started automatically labeling content its systems identify as AI-generated.
TikTok's Community Guidelines require disclosure for AI-generated content that depicts realistic scenes of people or events.
The EU AI Act Requirement (August 2026)
The EU AI Act's full compliance deadline is August 2, 2026. For a broader look at GDPR and the AI Act as they apply to video tools, see GDPR-compliant AI video tools: what to look for. For creators distributing content to EU audiences, one of the relevant requirements is transparency about AI-generated material: content that a reasonable person could mistake for authentic footage must be labeled as AI-generated.
The Act does not specify the exact format of the label, but the expectation is that it is prominent enough that viewers encountering the content would see it. A small note in a YouTube description that most viewers never open is unlikely to satisfy the intent of the regulation.
If you publish videos to EU audiences and they contain realistic AI-generated imagery or voice synthesis, a clear disclosure in the video itself (an on-screen label at the start, or a persistent watermark) is the safer approach than description-only disclosure.
A Practical Disclosure Format
For a video with AI narration and AI-assisted production but no realistic AI depictions of real people or events, a simple disclosure in the description is sufficient and appropriate:
This video uses AI-generated narration and AI-assisted production tools.For a video with AI-generated footage that depicts realistic scenes:
This video contains AI-generated imagery. Some scenes were created using
AI tools and do not represent real events or people.If you are cloning or using a synthetic voice that sounds like a specific person:
The narration in this video uses an AI voice model. The voice is
synthetic and does not represent the actual speech of [person/character name].Why Proactive Disclosure Is the Better Position
Aside from the legal and platform requirements, there is a practical argument for disclosing even when you technically do not have to.
Audience trust is harder to rebuild than it is to build. A viewer who discovers that a video they believed was authentic was actually AI-generated feels deceived, even if the content was accurate. A viewer who sees a disclosure upfront forms no false expectations.
For channels in finance, health, or any domain where credibility matters, proactive disclosure builds a reputation for honesty that distinguishes you from channels that are less forthcoming. The disclosure becomes a marker of quality rather than a flag of concern.
YouTube's enforcement history also suggests that the channels penalised are rarely the ones that disclosed proactively. Penalties concentrate on channels that tried to hide AI involvement in realistic or misleading content.