PROFESSOR OF TECHNOLOGY & SOCIETY · BANGOR UNIVERSITY
What happens when machines seem a bit alive?
I research technologies that read and simulate human emotion, from emotion recognition to AI companions, and how we should govern them. I direct the Emotional AI Lab and chaired IEEE 7014.1, the international standard on emulated empathy in AI.
What I argue
- 01
Chatbots are empathic by default, and most regulation hasn't noticed.
Large language models produce empathy-like behaviour as a side effect of how they're built. Rules written for “emotion recognition” miss it entirely.
- 02
The EU AI Act defines emotion recognition too narrowly.
Relabel “emotion” as “expression” or “cues”, or drop the biometric data, and a system can slip outside the definition.
- 03
Law moves slowly. Standards can get there first.
Standards bring experts together on issues that look fringe to the public but matter to specialists, before those issues become crises.
NEW BOOK · WITH VIAN BAKIR · BRISTOL UNIVERSITY PRESS · 2027
A Bit Alive
Should AI be pushing our social buttons?
Chatbots now listen, adapt and respond as if they know us. Sometimes they seem to care. Written for general readers, A Bit Alive looks at AI companions, ghostbots, AI influencers and the attachment economy behind them. It asks whose interests these systems really serve, and argues that none of this is fixed: the terms of our relationship with social AI are being set right now.
Ideas
- empathic median.
- Media and technologies that sense and interpret our feelings, moods and intentions. Introduced in Emotional AI: The Rise of Empathic Media (2018).
- hybrid ethicsn.
- An approach to technology ethics that starts from the fact that people are already entangled with technology, and asks what good terms for that relationship look like. From Automating Empathy (2023).
- emulated empathyn.
- Systems that simulate understanding of, and response to, our emotional states. The subject of IEEE 7014.1-2026.
Recent writing
All publications →FOR JOURNALISTS
Need a comment?
I can talk about AI that interacts with human emotion, AI companions, anthropomorphism and AI, and why technical standards matter.
Interviewed or quoted by BBC Radio 4, The Guardian, Financial Times, Wired, The Washington Post, MIT Technology Review, Scientific American
FOR RESEARCHERS
Collaborate
The Emotional AI Lab works across philosophy, law, media and social science, with partners in the UK, Europe, Asia and Nigeria.
FOR ORGANISATIONS
Consultancy
As AI moves from text to voice, video and persuasive emulated empathy, the risks grow. I help organisations understand what emotional AI means for their users, their risk exposure and the policy landscape ahead.
Standards & regulation interpretation · Biometric & sentiment privacy review · Threat modelling for voice-cloning, deepfake and synthetic-intimacy scams · Executive briefings · Investor due diligence · Expert witness
Interpretation and readiness, not certification against the standard I chair.