Vasilios is a Principal Research Scientist at the Alan Turing Institute, where he leads research on AI cyber risks.
His research focuses on how frontier AI systems shift offensive cyber capabilities, and how to evaluate and mitigate those risks in ways that are credible to both technical audiences and policymakers. Central to this work is measuring AI-enabled cyber threats, including through the AI Cyber Risk Benchmark, and leading the cyber offence section of the International AI Safety Report (2024-25 and 2025-26). Alongside empirical research, he contributes to technical AI governance, developing the tools and frameworks needed to support credible oversight of frontier models and strengthen cyber resilience in high-stakes systems.
