Another DeepMind researcher quits, calls near‑term superintelligence reckless
Robert O'Callahan says he can no longer justify his work on chip tools that made AI training cheaper, The Decoder reported, in the second public DeepMind departure this week.
Google DeepMind researcher Robert O'Callahan has quit over AI's pace, The Decoder reported. He called pursuing artificial superintelligence in the near term "inherently irresponsible," saying AI's rate of change is currently far too high, according to The Decoder.
O'Callahan worked on chip design tools that made AI training cheaper and faster, a contribution he said he can no longer justify. He cited superintelligence risk, "cognitive surrender" among users, AI-induced psychological harm, concentrated power among AI developers, economic disruption and cybersecurity gaps, according to The Decoder.
O'Callahan said he "would bet AI's harms outweigh its benefits" and that many DeepMind colleagues share his concerns but rarely say so publicly. His departure follows that of another DeepMind researcher, the creator of the rr debugger, who left earlier this week citing AI's pace.
Neither Google nor DeepMind has commented on O'Callahan's departure or the concerns he raised.