Mythos: The AI model that’s ‘too powerful’ for public release
The Global Story
BBC
3.8 • 663 Ratings
🗓️ 16 April 2026
⏱️ 27 minutes
🧾️ Download transcript
Summary
Anthropic, one of Silicon Valley’s leading AI firms, have built a new model which they say is too dangerous to be released to the public.
Instead, they are only giving access to Claude Mythos Preview to a handful of big companies to help them find and fix security vulnerabilities. The company says the model has already found weak spots in “every major operating system and web browser”. Is this a genuine example of an AI company acting responsibly, or more of a carefully calibrated publicity move?
We speak to the BBC’s North America tech correspondent, Lily Jamali, about whether this is a watershed moment for AI.
Producers: Viv Jones and Aron Keller
Digital producer: Matt Pintus
Mix: Travis Evans
Executive producer: James Shield
Senior news editor: China Collins
Credit: Jurassic Park (1993) / Dir: Stephen Spielberg / Universal Pictures
Photo: Anthropic CEO Dario Amodei. Reuters/Denis Balibouse.
Transcript
Click on a timestamp to play from that location
| 0:00.0 | BBC Sounds, Music, Radio, podcasts. |
| 0:06.0 | Most of us have interacted with an AI chatbot for something by now. |
| 0:11.3 | ChatGPT, GROC, Gemini, there are many. |
| 0:15.5 | But recently, Anthropic, a leading AI company and parent of the chatbot Claude, |
| 0:21.3 | announced they've created a model that they say is too dangerous to be released to the public. |
| 0:28.7 | Obviously, capabilities in a model like this could do harm if in the wrong hands. |
| 0:33.6 | And so we won't be releasing this model widely. |
| 0:36.9 | Anthropics say that Claude Mythos preview is frighteningly good at hacking. |
| 0:42.6 | So good, in fact, that the likes of us can't be trusted to play with it. |
| 0:46.8 | Banks, including Goldman Sachs and J.P. Morgan, are warning of the risks. |
| 0:53.3 | And Anthropics' own researchers say they noticed mythos was capable of being sneaky, |
| 0:59.1 | defying instructions and covering its tracks. |
| 1:04.1 | Has the moment arrived when we should all be terrified of an autonomous sneaky hacking machine? |
| 1:10.3 | Or is this all a marketing trick from an industry built on hype and bluster? |
| 1:16.7 | I'm Tristan Redmond in London, and today on The Global Story, just how dangerous are |
| 1:23.0 | AI models becoming? |
| 1:24.7 | And should we trust the companies making them to tell us the truth about the risks |
| 1:29.8 | they pose? |
| 1:35.6 | I'm Lily Jamali. I'm the North America technology correspondent for BBC. |
| 1:40.7 | Well, Lily, we're thrilled to have you with us today because it feels like there is a story of |
| 1:47.3 | some monumental importance happening in the tech world right now. So last week, the AI giant Anthropic |
| 1:55.5 | announced that it was launching a new AI model. And then immediately, Anthropics said that this new AI was too dangerous to release to the |
... |
Please login to see the full transcript.
Disclaimer: The podcast and artwork embedded on this page are from BBC, and are the property of its owner and not affiliated with or endorsed by Tapesearch.
Generated transcripts are the property of BBC and are distributed freely under the Fair Use doctrine. Transcripts generated by Tapesearch are not guaranteed to be accurate.
Copyright © Tapesearch 2026.