FAR.AI Selected to Lead EU AI Act CBRN Risk Consortium

We're pleased to share that FAR.AI has been selected by the European Commission's AI Office to conduct technical safety research supporting the implementation of the EU's landmark Artificial Intelligence Act. We'll tackle one of the most critical safety challenges posed by advanced AI systems: preventing misuse of AI systems to help produce Chemical, Biological, Radiological, and Nuclear (CBRN) threats. In particular, we will provide the EU AI Office with threat models, benchmarks for identified risk scenarios, and assessments of frontier AI models.

We will lead the consortium for CBRN Risk Modelling and Evaluation following the successful bid for “Artificial Intelligence Act: Technical Assistance for AI Safety” tender (EC-CNECT/2025/OP/0032, Lot 1). The consortium brings together leading expertise from FAR.AI (AI safety research and red-teaming), SecureBio (biological threat assessment), SaferAI (AI governance and risk modeling) with specialized expertise from subcontractors: GovAI (threat modeling), Nemesys Insights (chemical, biological, radiological, and nuclear threats), and Equistamp (evaluation engineering).

Background

The EU AI Act entered into force on August 1, 2024, establishing the world’s most comprehensive regulatory framework for artificial intelligence. The act aims to ensure transparency, safety, and accountability in the development and deployment of AI technologies. A core mandate of the Commission's new AI Office is evaluating general-purpose AI models classified as systemic risks under the AI Act—including assessing risks posed by lowering technical barriers to the production or misuse of chemical, biological, radiological, or nuclear (CBRN) agents or materials. 

What we’re doing

Over the next three years, FAR.AI will collaborate closely with the AI Office to develop an integrated risk assessment framework for CBRN threats. Together with our consortium partners, we will conduct:

• Risk modeling and scenario development: We’ll organise risk modelling workshops together with the AI Office, producing risk models, scenarios, and thresholds to inform further work on evaluations. 
• Development of evaluation tools: We’ll synthesize available evaluation work into the developed risk framework, and shape standardised reporting procedures for CBRN risks.
• Red-teaming and adversarial testing: We’ll conduct rigorous evaluations to stress-test AI safety and the extent to which AI developments might uplift bad actors’ CBRN capabilities.
• Risk monitoring: We’ll regularly brief the Commission on new developments (e.g. models, risk sources, elicitation techniques, mitigation, incidents) as the field evolves.

Additionally, FAR.AI will support Harmful Manipulation Risk Modelling Evaluation, conducting  red-teaming as subcontractors on another lot of the same tender, Lot 4.

About FAR.AI

FAR.AI is an independent nonprofit research organization dedicated to ensuring advanced AI systems are trustworthy, secure, and aligned with human values. Founded in 2022, we conduct foundational technical research on AI safety challenges while building the field through workshops, fellowships, and collaborative initiatives. Learn more at far.ai.

For inquiries, contact us at media@far.ai.