The EU can now fine makers of ChatGPT-style AI models: what the AI Act enforcement means

Since August 2, 2026, the European Commission's AI Office can enforce the AI Act's rules for general-purpose AI models, with fines of up to €15 million or 3% of global turnover.

Published October 25, 2026 9:02am EDT | Updated October 25, 2026 11:01am EDT · 2 min read

Add ScoviAI on Google

Flag of the European Union (public domain, Wikimedia Commons) (European Commission)· 2 min read

Key points

The AI Office's enforcement powers started on August 2, 2026
Fines for breaking the rules on general-purpose AI models: up to €15 million or 3% of global turnover
The AI Office can test models and order them restricted, withdrawn or recalled
Models launched before August 2, 2025 have until August 2, 2027

The European Union's AI Act has moved from rules on paper to rules with teeth. Since August 2, 2026, the European Commission's AI Office, together with national authorities, can enforce the law, including its rules for general-purpose AI models: the large models behind tools like ChatGPT, Claude and Gemini.

What counts as a general-purpose AI model?

The Commission's guidelines treat a model as general-purpose when it's trained with a very large amount of computing power (more than 10^23 operations) and can produce language, images from text, or video from text. In practice, that covers the major chatbot and image models.

What the AI Office can now do

Ask for information and documents.
Request access to a model to evaluate it.
Require risk-reduction measures.
Order a model to be restricted, withdrawn or recalled from the EU market.
Issue fines.

How big are the fines?

Breach	Maximum fine
Prohibited AI practices	€35 million or 7% of worldwide annual turnover
Other breaches, including general-purpose model rules	€15 million or 3% of worldwide annual turnover

In each case, the higher of the two amounts applies. Giving incorrect or misleading information to the AI Office can also be fined.

Is there a grace period?

Yes, for older models. Models already on the market before August 2, 2025 have until August 2, 2027 to comply. Newer models must comply now.

How companies comply

The main route is the General-Purpose AI Code of Practice, a voluntary set of commitments. The Commission says the AI Office works closely with companies that sign it. The Commission's guidelines aren't legally binding, but they show how it reads the law and will guide enforcement.

Can I report a problem?

Yes. The AI Office has a complaint tool for people and businesses, a whistleblower tool for people who work at AI companies, and a separate channel for businesses that build products on top of these models.

What it means for you

In the EU, the companies behind major AI models are now accountable to a regulator that can test their models and fine them.
Outside the EU, expect the effects anyway: big AI companies usually apply one set of practices worldwide.

So far, we haven't seen reports of fines being issued.

Sources

Commission starts enforcing AI Act rules and new transparency requirements on 2 August (European Commission)
The enforcement framework of the AI Act (European Commission)
Guidelines for providers of general-purpose AI models (European Commission)

Source: European Commission

ScoviAI Team is the ScoviAI newsroom. We write every article ourselves from the original announcements and link our sources. Spotted a mistake? Tell us and we'll fix it.