Research
My research connects cybersecurity, AI security, multi-agent systems and evolutionary game theory to study how intelligent systems behave when trust, incentives and adversarial pressure interact.
Security is a prerequisite for trustworthy AI, but trustworthiness extends beyond attack resistance. I study security, reliability, verification, responsible deployment and the conditions under which users and organisations can justifiably rely on intelligent systems.
AI security · adversarial AI · trust · verification · responsible AI
As AI moves from models that answer questions to agents that act and interact, new questions arise around delegation, cooperation, competition, deception, reputation and collusion. My interest is in designing environments where trustworthy behaviour remains strategically viable.
agentic AI · MAS · cooperation · dynamic trust · strategic behaviour
I use evolutionary and game-theoretic perspectives to reason about adaptive security: how attacker, defender and cooperative strategies change over time, and how incentives, commitment and information shape collective outcomes.
EGT · adaptive security · mechanism design · cooperation · commitment
Distributed learning creates benefits without centralising all data, but introduces strategic and adversarial participants. I am interested in poisoning, backdoors, collusion, trust management and secure cooperation among federated clients.
federated learning · poisoning · backdoors · collusion · client trust
AI is becoming part of both cyber offence and defence. I study the transition from static attack-defence models towards adaptive ecosystems in which intelligent attackers and defenders learn from one another.
AI-enabled defence · adaptive attackers · threat detection · autonomous response
Healthcare provides a high-stakes environment for studying trustworthy AI. My work includes multi-agent medical AI, secure collaborative learning, bias-aware reasoning and the security of AI-assisted decision processes.
medical AI · safety · bias · multi-agent reasoning · secure collaboration
Current direction
My current research direction increasingly focuses on autonomous and collaborative AI: how agents establish trust, how malicious behaviour spreads or is contained, how incentives affect cooperation, and how security mechanisms remain effective as agents learn and adapt.