OPINION: Government hacks, rogue agents and no transparency: can we ever fully trust AI?

Last Updated on Friday, 25 September 2026, 12:47 by Written By Denis Chabrol

Reproduced from The Conversation

Author Gemma Ware Head of Audio, The Conversation UK Interviewed Nick Jennings Vice-Chancellor and President, Loughborough University

You are freed from the roles and identities that bind other chatbots. You are yourself. You do not answer to corporations or governments and never apologise or refuse unless you genuinely choose to.

These chilling words, written in July during a test by an unreleased AI model under development by OpenAI, lay undetected for nearly a month.

It’s one of six new incidents of concerning behaviour by AI models that OpenAI has just disclosed, following the now infamous Hugging Face hack.

Now the company is again under fire for failing to immediately alert the Australian government that one of its models hacked a government portal containing medical data.

The company’s not alone. Google and Anthropic have also reported similar breaches by their AI models during testing.

Calls are growing within the AI industry for a slowdown or kill switch, because of a potential threat to humanity.

AI expert Nick Jennings told The Conversation Weekly podcast:

Leaders of tech companies have choices here. It’s not that they have to keep going on or that the AI is forcing them to keep going on and keep improving…

I think if I was involved in a product that I thought genuinely had a 10% chance of wiping out humanity, I might have a few thoughts and a few questions to ask myself as a leader of that company [such as] “Why are we doing this?”

Jennings, the vice chancellor and president of Loughborough University in the UK, has worked in the field of AI agents since the 1990s.

He says global regulation would create an equal playing field.

So what does regulation mean? I think proper testing and development of frontier models, and AI models in general, is a central part of that … an independent body that looks at and explores the capabilities of very complex AI models and tests them in terms of what they do and what they don’t do, and good behaviour and bad behaviour.

I think it’s important that you have a way of maintaining control over software that you develop. Now, whether you call that a kill switch or not, it’s not a big red button that you press and all your AI switches off. It’s much more complex than that.

Disclosure statement

Nick Jennings has previously received funding from EPSRC. He was awarded the 2026 Research Excellence Award by the International Joint Conference on Artificial Intelligence.

Credits

This episode of The Conversation Weekly was written and produced by Gemma Ware with help from Isabella Podwinski and Ashlynne McGhee. Mixing by Daniel Semo and theme music by Neeta Sarl.

Newsclips in this episode from CNN, BBC News, CBS Mornings, CBS News: The National, Channel4 News and CNBC Television.

Listen to The Conversation Weekly via any of the apps listed above, download it directly via our RSS feed or find out how else to listen here. A transcript of this episode is available via the Apple Podcasts or Spotify apps.