Is Regulation Enough to Stop AI Systems Killing Us All?
The Guardian
READ
Details
- Date Published
- 6 Oct 2026
- Priority Score
- 4
- Australian
- No
- Created
- 6 Oct 2026, 08:00 pm
Description
Letter: Martyn Thomas wonders how any rational person can argue that independent oversight and regulation will be sufficient
Summary
This contribution highlights the critical absence of rigorous safety cases for frontier AI models, contrasting current industry practices with the established engineering standards used for nuclear power and aviation. The author argues that without detailed risk analyses and evidence of hazard control, calls for independent oversight remain speculative and insufficient to prevent catastrophic outcomes. The piece emphasizes the specific danger of systems that developers themselves admit could pose existential risks to humanity, asserting that policy must be grounded in rigorous technical facts rather than industry fantasy. It advocates for a shift in governance from vague regulation to the high-assurance safety proofs required in other life-critical industries.
Body
Sam Altman, the CEO of OpenAI, speaking at the OpenAI DevDay 2026 conference. Photograph: Jeff Chiu/APView image in fullscreenSam Altman, the CEO of OpenAI, speaking at the OpenAI DevDay 2026 conference. Photograph: Jeff Chiu/APLettersIs regulation enough to stop AI systems killing us all?Martyn Thomas wonders how any rational person can argue that independent oversight and regulation will be sufficientOnce again a frontier AI model has been withdrawn because it fails safety tests (OpenAI scraps release of new model over safety concerns in internal testing, 28 September). There is the usual call for “independent oversight and regulation”, though never any explanation of how this would work or, most importantly, what evidence should persuade a regulator that an AI system is adequately safe.There is a long engineering tradition of regulating safety-critical software systems, such as those controlling aircraft or nuclear power stations; international standards require “safety cases” – rigorous analyses that show that the possible hazards have been identified and controlled and that the probability of an accident is extremely low.Typically, the safety case should show, with a probability of at least 99%, that an accident that could cause multiple fatalities will not occur more often than once in 1,000 years. It is extremely hard to provide evidence for such low probabilities, even for aircraft that at worst might kill 1,000 people in a single accident. The developers of frontier AI systems tell us that their systems might kill the whole of humanity, yet they have provided no detailed risk analyses, no safety cases that could be independently assessed, and no evidence that such safety cases could ever be produced for frontier AI systems.How, then, can any rational person argue that independent oversight and regulation will be sufficient? Policy must be based on rigorous analysis of the facts, not on fantasy.Martyn ThomasFellow and emeritus professor of IT, Gresham CollegeExplore more on these topicsAI (artificial intelligence)OpenAISam AltmanTechnology sectorTechnology policyComputinglettersShareReuse this content