The world’s biggest AI companies are taking “catastrophically high” risks in their race to develop superintelligent systems, a former OpenAI researcher has warned MPs, as Parliament examines whether tech giants can be trusted to regulate themselves.
Daniel Kokotajlo, who quit the ChatGPT maker over safety concerns, told the Business, Innovation, Science and Trade (BIST) Committee on Tuesday that the industry was developing increasingly powerful AI models without knowing how to control them.
“I would say that the risk tolerance across the industry is catastrophically high”, he told MPs.
The warning came as former researchers from OpenAI and Anthropic faced questions over whether the development of advanced AI was outpacing the safeguards designed to prevent the technology from behaving unpredictably.
Kokotajlo said he had resigned after losing confidence in OpenAI’s willingness to act responsibly as it pursued superintelligence, referring to systems capable of outperforming humans across virtually all tasks.
“I came to believe that OpenAI, in fact, would not behave responsibly around the creation of this technology”, he said.
He also alleged that staff had faced pressure to play down concerns about the technology.
“I did feel pressure to mince my words a bit in my public communications”, he said, adding that some internal research and forecasts was stopped from being published.
Steven Adler, another former OpenAI researcher, told MPs there had been resistance within the company to describing its technology as “dangerous”.
“Unfortunately, the reality of the technology – whether it is dangerous – does not depend on whether you call it dangerous”, he said.
Adler, who left OpenAI in late 2024, acknowledged that its internal culture may have changed since his departure.
OpenAI and Anthropic have been contacted for comment. Both firms have recently called for a slowdown in the development of the most advanced AI models to enable safety measures to keep pace.
AI safety fears mount as MPs scrutinise tech giants
Geoffrey Irving, a former researcher at OpenAI, Anthropic and Britain’s AI Security Institute, warned MPs that progress in AI capabilities was far outstripping advances in safety research.
“We are still undergoing very, very rapid capability advances without corresponding advances in safety”, he said.
Asked whether he stood by his assessment that advanced AI could have a 50 per cent chance of destroying humanity, Irving said a “coin flip is roughly the right model to have”.
The figure reflects his personal assessment of a hypothetical future risk, rather than an established scientific probability.
Adler urged governments to impose clearer limits on AI developers, warning that competition was encouraging companies to justify increasingly dangerous decisions based on what their rivals were doing.
“What I hope governments will do is they will help set the ceiling essentially on how much risk these companies can impose”, he said.
The hearing follows AI minister Kanishka Narayan’s warning on Monday that a major breach of Britain’s cyber defences was likely to be inevitable.
It also comes after Washington restored the UK’s access to advanced American AI models for pre-release safety testing, following a temporary restriction that raised questions about Britain’s ability to scrutinise the technology independently.
Anthropic itself warned on Monday that its models may soon be able to self-replicate and takeover devices like ovens and TVs. In filings for its upcoming IPO, the company reportedly said its technology could pose “existential risks to humanity”.
Executives from OpenAI, Anthropic, Meta and Google DeepMind are due to face MPs later on Tuesday, followed by senior government officials.