Sciencepulse

Forget the Terminator: The AI Risks Worth Worrying About Now

AI·November 15, 2026

Forget the Terminator: The AI Risks Worth Worrying About Now

Artificial intelligence has spent decades as a villain in fiction, from Mary Shelley's Frankenstein to the Terminator films. Many people still treat it as a technology of the future. In practice it is already built into everyday life, from the facial recognition that unlocks a phone to the translation apps people use on holiday.

That gap between how AI feels and how it works is the subject of a special 300th episode of the long-running interview series The Life Scientific, recorded before a live audience at London's Royal Institution. Presenter Jim al-Khalili speaks with Michael Wooldridge, professor of computer science at the University of Oxford and director of Foundational AI Research at the Alan Turing Institute. Wooldridge believes the most common fears about the technology are misplaced. The episode argues that the dramatic picture of machines turning on their makers pulls attention away from problems that are already doing damage, including deepfakes and electoral disinformation.

Wooldridge's aim is to demystify the field. He plans to use this year's Royal Institution Christmas Lectures to take the lid off modern AI and examine how far it could develop. The conversation also touches on more personal territory. He recalls a time when sending an email was a thrilling novelty, and he admits that people at parties often want to talk to him about the Terminator. In a lighter moment, he is challenged by ChatGPT to come up with a novel future use for AI.

The broader point is one that matters well beyond academia. Public debate about AI often swings between hype and dread, and both tend to crowd out plain explanations of what the systems can and cannot do. An expert who can describe the technology without either alarm or cheerleading is useful to anyone trying to decide how much attention, and how much regulation, the current tools deserve.

Reporting based on an external source.