European Commission's chief scientific advisors and ethics group warn against weakening EU AI rules
European Commission science and ethics advisors · Joint statement
Read the source (opens in a new tab)- What happened
- The European Commission's Group of Chief Scientific Advisors and its European Group on Ethics in Science and New Technologies published a joint statement on governing AI on November 16, 2026.
- Key points
- The groups say "Traditional guard rails will not suffice," and that enforcement can lack political will or capacity, "compounded by external efforts to delay or weaken" existing obligations.
- Deployment conditions should grow stricter as capability, autonomy and potential consequences rise; agentic systems may need limits on autonomous actions and a way to suspend or terminate them.
- The EU should build its own capacity to evaluate frontier models, including cyber dual-use and agentic capabilities, because "Companies should never be the sole arbiters" of their systems' safety.
- Significant incidents and near misses must be recorded and assessed, and evaluation methods must keep pace with capabilities rather than rely on benchmarks built for earlier systems.
- Evidence
- Joint statement from two independent advisory groups, published by the Commission's research directorate. The publication says its contents do not necessarily reflect the Commission's position.