P06 active
AI Safety, Evaluation & Governance
Ensuring intelligent systems are reliable, secure, measurable and accountable, through rigorous evaluation, guardrails and responsible AI practices.
Research Areas
- · AI evaluations and benchmarks
- · Agent security
- · Guardrails and policy enforcement
- · Interpretability and observability
- · Responsible AI and governance
Research Questions
- 01 How should autonomous agents be evaluated for safety and reliability?
- 02 What security vulnerabilities exist in tool-using agent systems?
- 03 How can AI systems be made more interpretable without sacrificing performance?
Research Projects
No research projects are currently assigned to this programme. Projects will appear here as the research agenda develops.
Publications
Publications from this programme will appear here as research projects progress through the YARA review and publication lifecycle.