What happened
On 5 November 2026 Capgemini announced it was selected by the European Commission's AI Office (DG CNECT) to lead a consortium — with Wavestone and France's LNE — supporting assessment of cyberthreats and cyber risks posed by advanced/general-purpose AI models under the EU AI Act, combining cyber-risk management, operational security, AI testing and conformity assessment. Contract value was not disclosed.
Why it matters
This is the AI Office standing up independent third-party capability to evaluate systemic-risk and cybersecurity obligations of frontier AI models under the AI Act — the practical enforcement instrument behind Articles 91/92 evaluations and systemic-risk designations. It makes concrete that EU enforcement of advanced-model safety and cyber-risk obligations is moving to externally supported conformity assessment.
Action needed
Providers of general-purpose and systemic-risk models should anticipate external, regulator-directed cyber-risk and safety assessment of advanced models in the EU, and bolster cyber-risk evidence, system cards and conformity documentation accordingly.