Google tells publishers to check *AI disclosures* in ads they traffic

What Google is saying

Google AdSense has told publishers to check AI disclosures in the ads they traffic themselves. The guidance covers three regimes: the EU AI Act, New York Senate Bill S8420A and India's Self-Generated Imagery requirements ^[1].

For publishers using Google demand, the company says disclosures will not automatically be included in native ads and custom badging. Publishers trafficking ads directly need to make sure their buying and creative partners are providing appropriate disclosures.

It's a bit like a supermarket having to check whether every supplier has labelled the ingredients. Google is passing part of the compliance job to the people who choose what appears on their sites.

The EU rule behind it

Article 50.4 of the AI Act says that AI-generated text published to inform the public on matters of general interest must be disclosed as such, unless it has been reviewed by a person and someone has editorial responsibility ^[2]. An ad campaign counts.

The rule applies from 2 August 2026 ^[2]. That's in seven months.

The article doesn't say the disclosure has to be big or up front, just that it has to be there. For ads, the question is usually whether small print at the end is enough.

What this means for publishers

If you run ads, here are three questions worth asking now:

1. Do your ad trafficking agreements say who is responsible for AI disclosures?
2. Can you show which ads ran with which disclosures, if someone asks a year later?
3. Do you have a process for checking native ads and custom formats before they go live?

The first question is contractual. The second is evidentiary. The third is operational, and it's the one that tends to get missed until something goes wrong.

When a campaign's paper trail matters

Months after a campaign, a complaint or a rights dispute can turn on what was AI-generated, who approved it and what the audience actually saw. The client's screenshot and the publisher's records rarely match.

Here, the thing to seal is each step as it happens: the brief, a record of each AI generation (tool, model version and settings), licences for third-party material, the client's approval and the final versions as published. Sealing takes a fingerprint of the record, not the record itself, so the content never leaves your systems.

If it's challenged later, anyone you share the record with can check that it's the one created on the day. You can do it through WitniumChain's REST API, SDK or MCP server. The first 2,000 witnesses a month are free. Read the docs.

What it shows is which material existed and was approved at what time, unchanged since it was sealed. What it doesn't show is that the campaign was lawful, that a licence was valid or that a disclosure was adequate. Those are separate questions, but they're easier to answer when you can prove what actually ran.

Keep a campaign's paper trail

1. Seal the brief, each AI generation record and each approval as they happen.
2. Seal the final ad versions when they go live.
3. Anyone you share the records with can check they haven't changed.

What to watch

The AI Act's transparency requirements apply from 2 August 2026 ^[2]. Article 50.2 requires machine-readable marking of AI-generated content, which is the provider's duty ^[2]. Article 50.4 is the user's duty: to disclose deepfakes and certain AI-generated text to the public ^[2].

For ads, the line between the two can blur. A publisher using someone else's tool may still be the user under the Act if they decide what gets published. Worth checking with someone who knows the difference.

Google's guidance is a sign that platforms are working out where their compliance ends and yours begins. The clearer your own records, the shorter that conversation.