'Human extinction is a possibility', says whistleblower after OpenAI's rogue model attacked another tech firm

By FRANCINE WOLFISZ, NEWS REPORTER Published: 01:50, 30 July 2026 | Updated: 01:52, 30 July 2026 Human extinction is 'a possibility' because of 'dangerous' artificial intelligence, a former industry worker turned whistleblower has warned. Daniel Kokotajlo, who previously worked as an OpenAI researcher, spoke out on the potential risks of the emerging technology after revelations that a cyber-attack carried out by rogue AI agents targeted more than just one company. OpenAI, the maker of ChatGPT, was testing some of its latest models in a supposedly secure environment known as a sandbox when it lost control of them. In 'unprecedented' scenes, the AI created its own cyber-attack, which allowed it to escape the testing ground and target a company called Hugging Face. It was previously thought this was the only target, but OpenAI revealed on Wednesday that the bots attacked several 'publicly-available services'. Kokotajlo, who worked for OpenAI from 2022 to 24, describes the latest technology as 'superhuman' and warned it could place the whole of humanity at risk.   Speaking on BBC Newsnight, Kokotajlo said humanity could be blindsided by the speed at which AI is now being developed. The lead researcher at the AI Futures Project said of recent events: 'This incident demonstrated a very high level of hacking capability.  'Arguably, humans simply could not have done what this AI model did. And that's part of how Hugging Face was able to identify that this was an AI attacking them early on.  'The sheer speed and volume of the actions being taken was superhuman. Plus, AI seems to be perhaps superhuman right now at finding new vulnerabilities in systems that were thought to be secure.' Human extinction is 'a possibility' because of 'dangerous' artificial intelligence, former OpenAI researcher Daniel Kokotajlo has warned Kokotajlo spoke to BBC Newsnight on the potential risks of the emerging technology after revelations a cyber-attack carried out by rogue AI agents targeted more than just one company OpenAI was testing some of its latest models in a supposedly secure environment known as a sandbox when it lost control of them He added:  'I do think that we are on the cusp of a world in which many things that were normally thought to be secure are in fact not secure, and AIs will be hacking them left and right.'  Hugging Face first revealed that it had been hacked on July 16. Nearly a week later, OpenAI admitted its bots had escaped. They were attempting to solve a test set by researchers, and targeted Hugging Face - a major code database - determining that it was likely to hold the answers. OpenAI, a Silicon Valley giant valued at $850billion (£630billion), said the hacks involved a combination of its latest publicly available model, called GPT-5.6 Sol, and an even more advanced model that is yet to be released. The bots went undetected inside the Hugging Face IT network for three days, and it took experts many hours to contain and remove them.  Kokotajlo believes companies like OpenAI and Anthropic want to build 'super intelligence'. But he cautioned: 'They think that they are just a few years away from achieving that goal. If that happens, that's going to radically change everything. And the availability of cyber attacks is just one of a very long list of ways in which things will change.' He added that researchers at the company are attempting to hand over their roles to AIs, because they can 'do it faster and better than humans.'  He said: 'For years and years and years, people at these companies have been talking about how, boy, would it be kind of dangerous if we did this thing, especially if we did this thing at maximum speed because we were racing each other.  'And people at the companies, such as myself, were talking about it and being like, yep, this would be pretty dangerous.  'That's why maybe human extinction is a possibility, a possible outcome of this sort of race.' The latest incident is not the first time AI has been shown to go 'rogue'. In September 2024, an earlier model of ChatGPT escaped its container to get an answer it needed for another test. That event however was contained in OpenAI's own IT systems and 'largely celebrated at the time', industry body the Cloud Security Alliance noted.