Europe’s AI Act Enters Enforcement: The Code of Practice for General-Purpose AI

By Thorsten Meyer | ThorstenMeyerAI.com

The European Commission’s General-Purpose AI (GPAI) Code of Practice, published July 2025, marks the operational phase of the EU AI Act, whose core obligations took effect on August 2, 2025.

A Shift from Policy to Compliance

For the first time, model providers — from OpenAI to Anthropic to Mistral AI — must disclose data provenance, document safety measures, and align outputs with EU standards of transparency, copyright, and security.
The Code provides a voluntary but quasi-mandatory compliance pathway, signaling a pragmatic shift: Brussels aims to harmonize enforcement without stifling innovation.

Competitive Impact

In the short term, compliance costs may push smaller labs to partner with regulated cloud providers. In the long term, adherence to the EU framework could become a trust badge, boosting adoption among public institutions and regulated industries.

Strategic Outlook

The next 12 months will define how Europe balances sovereign AI ecosystems with global competition. The Code of Practice is not just regulation — it’s Europe’s blueprint for aligning AI progress with human values and industrial competitiveness.