It applies to every upload
A disclosure duty framed around AI-generated content attaches to the individual piece of content a user sees, not to a general terms-of-service statement covering the platform as a whole.
Platforms & Enterprise
Every upload is a disclosure decision. You can't make it by hand.
Since August 2026, platforms hosting or deploying AI-generated content face disclosure obligations that apply upload by upload, not policy by policy. Provenance Radar checks Content Credentials, metadata, watermark signals, and AI-generation indicators on every file, so your trust & safety team gets consistent evidence without opening each one by hand.
Used through the web app today · API for automated pipelines is coming soon
Why this matters
The obligation is per file, not per policy
Manual review works when volume is small enough for a person to look at everything. Platform scale means thousands, or millions, of uploads arriving continuously, each one now carrying its own disclosure question.
It doesn't pause for review
Uploads keep arriving around the clock. A review queue that only clears during business hours falls further behind every day new content ships.
It breaks down without a standard
Ask different reviewers to judge the same file by eye, on different days, under different workloads, and you get different answers. At platform scale, that inconsistency becomes the risk.
The cost of ad hoc
Nothing happens. Until it does.
Without a consistent, evidence-based way to screen uploads, the gap between what a policy says and what actually happens to a given file tends to show up in two places.
- You can't explain how you handled it
- When legal, a regulator, or a journalist asks how AI-generated content gets disclosed on your platform, "it depends who reviewed it that day" isn't an answer anyone wants to give.
- Unflagged content reaches users
- Ad hoc manual review misses things at scale. Every file that slips through without evidence attached is a trust & safety incident waiting to surface.
Every
upload needs an answer, whether it's the first of the day or the millionth.
Zero
manual review capacity scales with upload volume the way headcount does — the gap only widens.
One
consistent framework applied the same way to every file, so the answer doesn't depend on who's on shift.
Global regulatory status
This isn't an EU-only rule
AI-content disclosure obligations are landing in one jurisdiction after another. Here's what's already in force, and what isn't yet.
China
In effectIn effect since 1 September 2025
The Measures for Labeling AI-Generated Content require both a visible label and an embedded metadata label on AI-generated text, images, audio, and video.
South Korea
In effectIn effect since 22 January 2026
The AI Basic Act requires a visible label on realistic AI-generated content and an invisible watermark on stylised generated content. Penalties are deferred during a grace period, except for cases of serious harm.
European Union
In effectIn effect since 2 August 2026
Article 50 of the AI Act requires generative AI providers to mark synthetic content as machine-readable, requires deployers to disclose deepfakes, and requires AI-generated public-interest text to be labelled.
United States
Active, state by stateExpanding since August 2026
California's AI Transparency Act took effect 2 August 2026, requiring generative AI providers to offer watermarking and disclosure tools. Most states now have deepfake disclosure laws for political content — there is no single federal standard.
Australia
PendingNo dedicated law yet
Proposed mandatory AI guardrails were shelved in the December 2025 National AI Plan in favour of existing privacy and consumer law. The only date on the books is a narrow automated-decision disclosure duty from 10 December 2026 — not a general AI-content labelling rule.
This is a general summary for orientation, not legal advice — confirm current requirements for your jurisdiction with counsel before relying on it.
How it fits your pipeline
Built for the way platforms operate
Today, Provenance Radar is used through the web application, one file or one batch at a time. We're building an API so this can plug directly into your existing upload pipeline and run on content automatically, without a person in the loop for every file.
1
Content reaches your platform
A user uploads a document, image, or other file the way they always have. Nothing about your upload flow needs to change today.
2
It's checked for provenance evidence
Content Credentials, metadata, watermark signals, and AI-generation indicators are analyzed and classified by certainty, verified, declared, detected, inferred, or unknown, never collapsed into a single score.
3
Evidence reaches your team
A structured report is ready for your trust & safety queue, or for answering the question of how a specific file was screened, whenever that question comes up.
Illustrative pipeline. No API endpoint exists yet, automatic, API-driven scanning of every upload is on our roadmap, not available today.
Questions from platform teams
What's live today, what's on the roadmap, and what a report can and can't tell you.
Not yet. We're building an API so you can submit content programmatically and get structured evidence back on every upload. It hasn't shipped — today, Provenance Radar is used through the web application.
Give your team an answer that scales
Check Content Credentials, metadata, watermark signals, and AI-generation indicators on every file, so you have consistent evidence when it's time to explain how your platform handles disclosure.