Essex expert sounds urgent call on AI and human judgement

A University of Essex cyber expert is set to challenge one of the biggest blind spots in artificial intelligence: human judgement.

Dr Carolyn Swinney, an Executive Fellow in the School of Computer Science and Electronic Engineering (CSEE) and a Royal Air Force Cyberspace Officer, will deliver a headline keynote at Cyber Science 2026, one of the UK's leading cybersecurity research conferences, held at Royal Holloway, University of London (3-5 June).

Her message is bold and timely. As artificial intelligence becomes embedded in high-stakes decision-making, the greatest risk may not lie in the systems themselves, but in how they shape the thinking and behaviour of the people who rely on them.

In her keynote, The Threat We're Not Modelling: Securing Human Judgement in the Age of AI, Dr Swinney will argue that human judgement must be recognised as a core element of security, not an afterthought.

"Human judgement is not a soft skill; it is a security control," she said. "We must understand what AI is doing to the people making critical decisions."

Alongside her keynote, Dr Swinney will present new research exploring why AI systems can appear confident and authoritative even when they are wrong. Her work highlights a critical distinction between human and machine learning: while humans learn through experience and consequence, AI systems can repeat the same mistake with confidence, without the same form of reflection or adaptation.

"AI can repeat the same confident mistake endlessly without learning in the way humans do," she explained.

The findings point to a growing risk of overconfidence, misplaced trust, and flawed judgement, particularly in environments where decisions carry significant consequences.

Professor Reinhold Scherer, Head of the School of Computer Science and Electronic Engineering at the University of Essex, emphasised the wider significance of this work, noting that the underlying challenge is not new, but is taking on new urgency as AI becomes more pervasive.

"This challenge goes far beyond cybersecurity. It is about strengthening human agency in an AI-driven world. Knowing, understanding, and critical thinking are essential, and AI should amplify these strengths to help us make better, more responsible decisions."

At Essex, this is more than a message, it is a mission. CSEE is driving forward a vibrant cybersecurity research agenda through its dedicated research group, while investing in the next generation of talent. The School launched a new BSc (Hons) Cyber Security and offers a cutting-edge MSc Cyber Security, equipping students with the skills, mindset, and critical awareness needed to secure both systems and the people who depend on them.

Cyber Science 2026 brings together leading voices from academia, industry, and defence to explore the future of cybersecurity.

Dr Swinney's message is clear: as AI systems grow in capability, safeguarding human judgement will be essential to ensuring those systems are used wisely, effectively, and responsibly.