Michael I. Jordan calls AI doom warnings science fiction

By Splitfeed EditorialLast updated Oct 13, 2026, 17:21 UTC

Michael I. Jordan, whom Science magazine has previously described as the most influential computer scientist, made a pointed public intervention against what he characterized as alarmist AI rhetoric. He dismissed warnings about superintelligence as science fiction, described young AI researchers who have resigned over safety concerns as staging a grand moralizing performance, and labeled calls from AI companies for regulation as theater. His sharpest technical criticism was directed at Geoffrey Hinton's argument that the human brain can be reproduced in silicon and then scaled into superintelligence, with Jordan stating that none of the steps in that argument are credible.

The intervention matters because Jordan is not a peripheral figure delivering a contrarian take from outside the field. His standing in machine learning research gives his dismissal a weight that is difficult to set aside on credentials alone. The argument he is entering has real institutional stakes: several prominent AI labs have built safety teams, researchers have left positions citing existential concerns, and some company executives have publicly asked governments to impose oversight partly on grounds of catastrophic risk. Jordan is challenging the empirical and logical foundations of that entire framing, not merely its policy conclusions. The genuine dispute is whether the brain-to-silicon analogy is a plausible extrapolation from current machine learning or a category error, and that is a technical question on which credentialed experts hold strongly opposing views.

Because no camp reactions have been published yet, what each would typically argue can only be anticipated. The Pro-AI camp would likely welcome Jordan's comments as expert validation that doom scenarios are distracting from productive development and that regulatory calls from companies may serve competitive rather than safety interests. The Anti-AI camp would typically counter that dismissing risk as science fiction is itself a performance, and that Jordan's critique of Hinton does not resolve the broader question of whether advanced systems pose dangers that deserve serious institutional attention. The Middle Ground camp would generally be expected to treat Jordan's technical objections as worth engaging while resisting the conclusion that safety concerns deserve no structural response at all.

The clearest near-term signal to watch is whether Hinton or other researchers who have publicly endorsed the brain-to-silicon scaling argument respond directly to Jordan's specific claim that none of its steps are credible. A detailed technical exchange between figures of that standing would sharpen what is currently a high-level disagreement into something that could be evaluated on concrete grounds.