Researcher in AI, AI Ethics, AI safety, artificial moral agents, and interpretability. Imperial College London.
I have worked on AI since 2014, focusing on value-aligned moral agents, explicit logic-based moral reasoning, interpretability, and meta-decision-making for AI. My research keeps returning to a single problem: we do not yet have good top-down, logic-based, explicitly moral decision-making frameworks that a model can reason through. Nearly everything we use is bottom-up. Closing that gap, in my view, is one of the most consequential open questions in AI safety.
I lecture on Contemporary Philosophy and AI at Imperial College London, and previously designed and led the Department of Computing's first AI and Ethics course (2019–2023, 95.45/100 student satisfaction). I am AI & Ethics Lead on two Horizon Europe consortia in personalised cancer prevention (4P-CAN, €5.3M) and early diagnosis (FH-EARLY), and am establishing an Imperial-based research institute on AI in wellbeing and health.
Artificial moral agents and value alignment. Extracting and formalising the implicit value structure of AI systems — what I call moral paradigms — through structured, logic-based probing. Interpretability and the interpretation problem. Meta-decision-making for AI (Relevance, Representation, Reasoning). Persona-based evaluation of frontier model behaviour.
Designed, created, and led the first AI and Ethics course in Imperial College London's Department of Computing, 2019–2023. Student satisfaction 95.45/100. The course seeded multiple PhD collaborations and joint publications. Currently lecture on Contemporary Philosophy and AI at Imperial.
Email: cos@ethicos.co.uk
Based in London. Open to collaboration and conversation on AI, AI Ethics, AI safety, alignment, interpretability, and moral reasoning in AI.