Tommy is a researcher, technologist and policy advisor specialising in frontier AI safety and security. He is currently a Senior AI Policy Manager at the Centre for Long Term Resilience. Previously he led the development of a £10m AI platform for the UK government, which deployed novel ML techniques to analyse online disinformation. Prior to this, he was Head of Policy at the non-profit First Draft, where he led and consulted on projects with WHO, Stanford University, Unicef, UN, Google and Partnership on AI. He is currently completing a PhD on AI safety incidents, and how they have transformed the development of AI systems from 2012 to the present day.
Research
I have conducted a wide range of research specialising in AI incidents and disinformation, including in leading peer-reviewed academic journals. This has included threat assessments, threat actor uplift studies, and research into elections using digital methods and data science methods.
AI incidents and 'networked trouble': The case for a research agenda, Arxiv, 2024
The near-term impact of AI on disinformation, CLTR, 2024
The rise of “gaslighting”: debates about disinformation on Twitter and 4chan, and the possibility of a “good echo chamber”, Popular Communication, 20:3, 178-192
The Semiotics of Authenticity: Indexicality in Donald Trump’s Tweets. Social Media + Society, 4(3). https://doi.org/10.1177/2056305118800315
Diagram from my research into disinformation threat actor uplifts from LLMs based on a meta-analysis of over 70 research studies and expert interviews. (The near-term impact of AI on disinformation, CLTR, 2024).
Policy work
-

Preparing for AI security incidents
In a major report, I raised the alarm about the increasing likelihood of a major AI security incident, and developed 34 concrete policy recommendations for how the UK Government could pioneer a holistic AI security strategy. My report was covered in TIME Magazine, and I engaged Ministers and Special Advisors on my recommendations.
-

Incident reporting for AI security
I developed policy recommendations for how the UK should launch an incident reporting regime for AI incidents. I worked closely with UK government officials on operationalising these recommendations, and my report drove a national conversation about incident reporting with coverage in national news.