I work on AI assurance and civilisational resilience: building the technical foundations for independently verifiable claims about the behaviour of AI systems and the infrastructure they run on — from secure silicon to software to multi-stakeholder coordination.
Currently, I serve as Technical Specialist on Safeguarded AI, a ~£60M R&D programme at the UK's Advanced Research and Invention Agency (ARIA), where we're building a mathematical assurance toolkit that lets fleets of AI agents produce formally verified artifacts at unprecedented speed and scale.
Before ARIA, I co-founded and led Principles of Intelligence (formerly PIBBSS), a research organisation facilitating knowledge transfer from interdisciplinary sciences into AI safety. I've also been a Research Affiliate with the Alignment of Complex Systems research group, and a Research Manager at the Future of Humanity Institute, University of Oxford.
Selected Work
- Gradual Paths to Collective Flourishing. Essay exploring a vision for successfully navigating the AI transition through enabling stable, win-win coalitions between humans and AI systems. January 2026.
- AI Resilience. Research initiative and web resource on R&D priorities for civilisational resilience in the age of AI. September 2025.
- Faster AI Diffusion Through Hardware-Based Verification. Essay as part of IFP's Launch Sequence, identifying concrete but ambitious AI projects to accelerate progress in science and security. August 2025.
- Flexible Hardware-Enabled Guarantees. Co-authored framework and report series on leveraging hardware architecture for AI assurance. February 2025.
- Gradual Disempowerment. Co-authored analysis of incremental AI capability growth and structural risks to human oversight. January 2025.
- (Talk) "Gradual Disempowerment." Talk at Cooperative AI Summer School. (2025).
- Avoiding an AI Arms Race with Assurance Technologies. Policy analysis for AI Frontiers. Argues that assurance technologies – such as hardware-based mechanisms and others – can help temper an international AI arms race. January 2025.
- Towards Guaranteed Safe AI: A Framework for Ensuring Robust and Reliable AI Systems. Multi-author position paper defining a family of approaches for AI with quantitative safety guarantees, combining world models, safety specifications, and verifiers. (2024).
- The Value Change Problem. MA dissertation, organised as post series. Argues that AI alignment must address the problem of (il)legitimate value change — ensuring AI systems neither manipulate human values nor foreclose their legitimate evolution. 2023.