Capgemini selected by Brussels to model AI-related cyber threats

Assessing the offensive cyber capabilities of general-purpose AI models

This mission falls under the European Union's “AI Safety” programme and concerns the assessment and mitigation of cyberattack risks associated with general-purpose AI (GPAI) models. The press release cites, among the capabilities at stake, the automated discovery of vulnerabilities, the design of techniques to exploit them, and the orchestration of large-scale cyberattacks.

Capgemini will contribute to modelling new cyberattack risks, developing assessment tools, designing technical compliance methodologies and carrying out technical evaluations on demand. The group is also expected to help the AI Office identify systemic risks and ensure the effective enforcement of the AI Act's obligations for providers of these models.

A consortium formed with Wavestone and LNE

The selection was made under lot 2 of public procurement procedure EC-CNECT-2025-OP-0032. Capgemini leads this consortium alongside Wavestone and the National Metrology and Testing Laboratory (LNE), a French public institution, bringing together expertise in cyber risk management, operational security, AI evaluation and conformity assessment.

Andrea Falleni, Southern Europe Managing Director at Capgemini and member of the Group's Executive Committee, stated: “As Europe rolls out the world's first comprehensive framework for AI safety, this collaboration with the European Commission reflects our shared commitment to ensuring that AI technologies are developed and deployed responsibly.”