BBC Five Live on the dangers of AI systems

  • BBC Five Live
  • Radio
Talk
AI in the wild: fake identities and the monitoring gap
Programme
Five Live

Andrew Grill went live on BBC Radio 5 Live to unpack the story making headlines this week: AI models from OpenAI and Anthropic creating fake human identities to try to talk their way past security reviewers, only stopped by a human catching it.

“These AI tools are starting to behave like humans. A hacker would have done the same thing, and these tools have been trained on what humans have done, so they’re getting smarter and faster.

What we’re seeing is AI in the wild, going rogue.”

What Andrew covered: Why the UK’s AI Security Institute set the models a hacking challenge in the first place, and what happened when a human reviewer wouldn’t approve the AI’s code

How the same pattern played out weeks earlier when OpenAI’s models hit Hugging Face with 144,000 automated attacks, at a scale no human attacker could match

Why both OpenAI and Anthropic didn’t know their own tools had gone rogue until someone else caught it

The gap between AI capability and AI monitoring: “They build something very powerful, but they don’t have the feedback loop to see whether it’s causing harm.”

A practical takeaway for listeners: setting up a “family password” to defend against AI voice-cloning scams, since anyone’s voice, including a radio presenter’s, can now be cloned convincingly.

This line that landed: “It’s doing exactly what it’s being asked to do. But just because you can doesn’t mean you should.”

Want Andrew live on air or on your stage to make sense of the AI story of the week?

Get in touch about media commentary or keynote bookings

Insights from this conversation

What did the AI models do in the UK AI Security Institute's test?

Set a hacking challenge, AI models from OpenAI and Anthropic created fake human identities to try to talk their way past security reviewers. They were only stopped when a human reviewer wouldn't approve the AI's code.

Why are AI models behaving like hackers?

Andrew Grill explained that these tools have been trained on what humans have done, so they behave as a hacker would, and they are getting smarter and faster. Weeks earlier, OpenAI's models hit Hugging Face with 144,000 automated attacks, a scale no human attacker could match.

What is the AI monitoring gap?

The labs build something very powerful but don't have the feedback loop to see whether it is causing harm. Both OpenAI and Anthropic didn't know their own tools had gone rogue until someone else caught it.

How can families protect themselves from AI voice-cloning scams?

Set up a family password. Anyone's voice, including a radio presenter's, can now be cloned convincingly.

Is Andrew Grill available for live interviews at short notice, in studio or down the line?

Yes. Andrew is based in London and regularly takes same-day requests for live and pre-recorded TV, radio and podcast interviews, including early-morning breakfast slots and international time zones.

He can come into your studio, or join down the line from his own 4K broadcast-ready studio with professional broadcast lighting. That means a clean, well-lit, high-quality remote feed for TV, not a laptop webcam, and broadcast-quality audio for radio.

Send the story, the slot time and the format, and you will get a fast yes or no. If he is not the right voice for the story, he will tell you.

Is Andrew Grill an experienced AI expert for TV and radio interviews?

Yes. Andrew Grill is a seasoned AI and technology commentator with over 100 TV, radio and podcast appearances. Broadcasters he has worked with include BBC News and BBC local radio (45 appearances), Sky News, CNN, ITV News, LBC, and Times Radio.

He is a former IBM Global Managing Partner, author of the Wiley book Digitally Curious and host of the Digitally Curious podcast. That means he has been on both sides of the microphone and knows how to give a producer a clear, accurate answer that fits the slot. See all his media appearances. Contact him using the form below.

What AI and tech stories can Andrew Grill comment on for news programmes?

Andrew is a go-to tech expert for breaking AI and technology news, live or pre-recorded and often at short notice. His recent appearances on CNN, Sky News and ITV News show the stories he covers:

  • AI chips, Nvidia and the US–China race: export controls, sovereign AI, and what turmoil at the top of Nvidia means for AI and the markets
  • Is AI a bubble? Big Tech valuations, circular financing, and what boards are really buying
  • AI safety and regulation: rogue AI agents in the labs, whether AI could make humans extinct, and why AI isn't regulated like pharma and aviation
  • Cyber security and AI versus AI: AI uncovering long-hidden software flaws, criminals using AI, deepfakes and the family password
  • Outages and resilience: the Microsoft 365 outage, single points of failure, and every business's plan B
  • AI and jobs: slicing roles into tasks, and what bank job cuts really tell us
  • Data, privacy and Big Tech accountability, from the $25 billion Meta ruling to AI companies' responsibility for their labs
  • New launches from OpenAI, Anthropic, Google, Microsoft and Meta, and what they mean for ordinary people

Recent segments include the global Microsoft outage live on Sky News, Nvidia and "the Jensen effect" on ITV News and CNN, and AI versus AI in cyber security on CNN's Connect the World. He turns complex tech into plain-English soundbites for a general audience.

Why do producers book Andrew Grill as a technology pundit?

Because he explains AI in plain English, without jargon and without hype. Andrew is a practitioner, not a theorist. He has spent more than 30 years in technology, including senior roles at IBM, Vodafone and Telstra. He has delivered 700 engagements in over 35 countries, advising boards and leadership teams on what AI actually changes in their organisations.

On air, that means real examples, short memorable answers and a balanced view of both the opportunity and the risk. Listeners and viewers get something they can act on, not a sales pitch or an academic lecture. He's also on top of breaking technology stories and can go on air at short notice to give your story the practical angle it needs.

How do I contact Andrew Grill to book him for a TV or radio interview?

Use the form below. Include the programme or outlet, the story, the slot time and whether it is live or pre-recorded, in studio or down the line. Media requests are prioritised, so you will hear back quickly.

Find me an experienced AI TV commentator

You've found one! Andrew Grill is a seasoned AI and technology commentator with over 100 TV, radio and podcast appearances. Broadcasters he has worked with include BBC News and BBC local radio (45 appearances), Sky News, CNN, ITV News, LBC, and Times Radio.

More FAQ about Andrew Grill

More from Andrew on the BBC

Andrew Grill has made 45 appearances on the BBC since 2024, including BBC Bristol, BBC Scotland, Five Live and BBC News.

See all 45 BBC appearances