EU formally begins enforcement
against frontier AI companies
The EU AI Office will formally exercise
its enforcement powers starting August 2, overseeing the most advanced
general-purpose AI models.
According to the Associated Press, the office has added 38 staff members, and both OpenAI and DeepSeek are within the scope of monitoring.
Regulators can request documentation, require model evaluations, and order risk reductions. If compliance discussions prove ineffective,
they may impose fines of up to 3% of global annual revenue, or require restriction, withdrawal, or recall of the models.
Transparency rules—such as chatbots disclosing their AI identity and adding machine-readable watermarks to generated content—
will also take effect at the same time.
#
AI
##
TechNews
##
EUAIAct#