- Published on
AI isn’t going to end humanity ... right? – podcast
- Authors
- Name
- DeepHorus
A high-profile AI developer warned last week there is roughly a 10% chance artificial intelligence will wipe out humanity within the next decade. The claim came during a surge of warnings and resignations across the tech industry. The Guardian’s Today in Focus podcast examined the question on Monday 14 September 2026. Host Annie Kelly asked UK technology editor Robert Booth how worried the public should be.
For years AI safety debates stayed inside AI labs and academic circles. Recently, however, tech leaders have publicly resigned over existential risks. Some companies have also reportedly lost control of their bots. This has pushed the issue into mainstream politics and media.
The 10% figure was described as “more or less” the risk by the developer. A string of similar warnings has emerged over recent months. These include high-profile resignations and incidents where AI systems behaved unpredictably. Booth called it a “crossover moment” for the AI safety story.
The shift matters because public and political leaders are now paying attention. AI is no longer just a technical curiosity. It touches jobs, security, and democratic oversight. If the risk is real, regulators may need to act before systems become more powerful.
Booth said the AI safety story has long been discussed inside the AI world. Now, he noted, the general public and political class have become increasingly concerned. That broader awareness could pressure companies to prioritize safety. It might also lead to new laws and international cooperation.
The podcast did not predict doom, but it refused easy reassurance. Listeners are left weighing a small but catastrophic probability against the benefits of AI. The coming decade will test whether warnings translate into safer design. For now, the question “AI isn’t going to end humanity … right?” remains open.
