Skip to content

Topic

AI content labeling and provenance

Marking generated or manipulated media so audiences can tell what they are looking at, including icons, detection signals, and creation-time metadata.

Current stories

productConfirmed20 publishers

YouTube Music lets partners declare AI use in three tiers and adds its own detection

YouTube Music now tags tracks Fully, Partly or No Gen AI, drawing on partner declarations and its own detection tools. Partners that skip disclosure can still be labeled, The Verge reported, because YouTube may use "other signals."

Perspective Coverage

22 publishers
Builder
Builder 10%
Operator
Operator 59%
Investor
Investor 31%

Reality

Evidence72
Adoption
Not scored
Hype gap+6
Incentives40
Confidence78
productOne report1 publisher

YouTube Music will label AI tracks while Deezer promises a way to switch them off

Deezer will add a toggle before the end of 2026 that hides AI-generated music, while YouTube Music will only label tracks made primarily with generative AI. The toggle will lean on a detector that Deezer says could be weak to audio manipulation and unknown generation models.

Publishers:techradar.com

Reality

Evidence35
Adoption20
Hype gap+25
Incentives60
Confidence50
buildOne report1 publisher

Public SynthID Detector makes AI-label wording a design risk

Google made SynthID Detector publicly available worldwide in English on October 7, 2026, covering images, video and audio. According to The Neuron, studies find people sometimes rate accurate content as less credible once it is labeled AI-generated.

Publishers:theneuron.ai

Reality

Evidence55
Adoption30
Hype gap+10
Incentives
Not scored
Confidence50
productOne report1 publisher

Meta left a politician deepfake up because no trusted partner had flagged it

In the Oversight Board's two new decisions, a takedown waited on a partner organization's flag and an AI label waited on the uploader's own disclosure. Meta has 60 days to answer the recommendations, one of which asks for faster review.

Publishers:engadget.com

Reality

Evidence60
Adoption20
Hype gap+10
Incentives55
Confidence62
investOne report1 publisher

China's top court hangs AI deepfake liability on the subject's consent

The Supreme People's Court says an AI double made or shared without the subject's agreement is an infringement, and providers share the liability once they have been notified. Likeness clearance becomes a budget line, not a lawsuit.

Reality

Evidence40
Adoption55
Hype gap+18
Incentives62
Confidence38
productOne report1 publisher

Spotify Cleaned Up the AI Anime Takeover of Jazz Profiles After Futurism's Report

Four named jazz musicians had synthetic tracks and anime cover art attached to their catalog pages on three streaming services, and the labeling feature Spotify has been promoting would not have stopped any of it.

Publishers:dexerto.com

Reality

Evidence30
Adoption35
Hype gap+40
Incentives62
Confidence38
productOne report1 publisher

Instagram is stamping AI Content on photos that only had their backgrounds removed

Users say Meta's label lands on ordinary retouching while genuinely generated pictures go untagged, which leaves brands publishing edited photography to explain a badge nobody at Meta will explain to them.

Publishers:theverge.com

Reality

Evidence55
Adoption45
Hype gap+20
Incentives55
Confidence50
buildOne report1 publisher

The Commission's marking code turns AI disclosure into a per-asset classification problem

The code covers labeling and detection of AI output, including deepfakes, and points deployers at EU icons. The work that lands on engineers is deciding which assets count as materially manipulated, and recording that at creation.

Publishers:dev.to

Reality

Evidence24
Adoption
Not scored
Hype gap+22
Incentives76
Confidence34