Nick Clegg dismisses 'godlike AI' extinction fears as tech industry 'fumes'

Nick Clegg has dismissed fears that artificial intelligence could acquire a "godlike power to exterminate humanity," arguing that tech bosses spreading such scenarios are "breathing their own fumes."

The former UK deputy prime minister, who now works in the tech industry, said the sector's leadership is "winding themselves up into a lather" over hypothetical existential risks while paying insufficient attention to concrete harms that already exist.

Clegg's remarks, reported by The Guardian on September 24, 2026, cut against a strand of AI discourse that has dominated headlines for years. Since the launch of large language models into mainstream use, senior figures inside major AI labs have repeatedly framed superintelligent systems as a potential extinction-level threat to humanity. Clegg, who moved from Westminster to a senior tech-industry role after leaving politics, is calling that framing misplaced.

Known threats over hypothetical ones

At the core of Clegg's argument is a simple distinction: between risks that can be named, measured and regulated today, and speculative scenarios about future machine capabilities that no one can currently demonstrate.

Tech bosses, he said, should concentrate on "specific threats" — he named cybersecurity, bio-weapons and emotional dependence — rather than what he described as the "slightly hand-wavy view that this technology is unavoidably going to develop some godlike power which is going to turn on us and exterminate humanity."

That list is notable for what it contains. Cybersecurity and bio-weapons are long-standing policy concerns with existing regulatory frameworks and government attention. Emotional dependence — the psychological attachment users can form toward AI systems — is a newer worry, and one that speaks to current consumer products rather than hypothetical future models. Clegg is effectively arguing that the industry's risk conversation should be anchored in the present tense.

The language he chose is pointed. "Breathing their own fumes" suggests that doom-laden predictions are circulating within the industry and being mistaken for analysis — an echo chamber in which executives amplify each other's worst-case narratives. "Winding themselves up into a lather" makes the same point in more colloquial terms: the panic, in his view, is manufactured rather than evidence-based.

Why the argument matters

Clegg's intervention lands in a debate with real regulatory and market consequences. How governments frame AI risk determines what they regulate, how heavily, and at whose expense. An emphasis on long-term existential risk tends to justify sweeping safety regimes and, critics argue, can conveniently advantage the largest incumbent labs — the very companies most vocal about frontier risks — by raising barriers to entry.

An emphasis on present-day harms points regulators toward narrower, more immediate targets: securing critical infrastructure against AI-enabled cyberattacks, controlling access to biologically dangerous capabilities, and protecting users — particularly younger ones — from forming unhealthy dependencies on chatbot products. These are areas where legislation and enforcement can act on known problems rather than forecast ones.

The timing also matters. Clegg made his comments amid an active UK political news cycle, and his unusual career trajectory — from deputy prime minister to senior tech-industry figure — gives him a foot in both camps. When he says the sector is overstating its own existential stakes, the criticism comes from someone with insider knowledge of how these companies talk about themselves.

The counterweights he is arguing against

Clegg is not dismissing AI risk altogether. His position, as expressed, accepts that the technology carries serious dangers; he simply insists they are the dangers we can already see. That places him in a camp of researchers and policymakers who have long argued that measurable harms — bias, misuse, security vulnerabilities, psychological effects — deserve more attention than abstract superintelligence scenarios.

The opposing camp, which includes several prominent AI lab leaders, holds that if machine capabilities continue to advance, the potential loss of human control becomes a legitimate near-term planning concern rather than science fiction. Clegg's response — "hand-wavy" — is a direct rejection of that reasoning's evidentiary basis as he sees it.

The substance of his case rests on a burden-of-proof argument: those claiming AI will inevitably develop catastrophic autonomy should demonstrate the mechanism, not assume it. Until then, he suggests, resources and political attention are better spent on the three threats he explicitly named.

Clegg's framing is likely to sharpen an already polarized debate about where AI safety resources should flow. If his view gains traction among policymakers, expect regulatory pressure to shift further toward cybersecurity, biosecurity and product-safety questions — and away from the extinction-risk agenda that some of the industry's most prominent voices have spent years promoting.