Google DeepMind Researcher Quits, Calls Superintelligent AI Push «Inherently Irresponsible»
Robert O'Callahan, a researcher at Google DeepMind, has resigned, saying the current rate of AI change is far too high and that building superintelligent AI soon is inherently irresponsible. He worked on chip design tools that made AI cheaper and faster, a contribution he now says he can no longer justify.
A researcher at Google DeepMind has resigned, warning that the rapid development of artificial intelligence is outpacing society's ability to manage it. Robert O'Callahan, who worked on chip design tools that helped make AI systems cheaper and faster, said he can no longer justify contributing to that acceleration.
In his resignation, O'Callahan described the current rate of change in AI as «far too high» and argued that building superintelligent AI in the near term is «inherently irresponsible.» His work focused on tools used in chip design, a technical field that has indirectly lowered the cost of training and running large AI models. By improving those tools, he helped speed up the very progress he now views with alarm.
O'Callahan said many of his colleagues share his concerns but rarely speak out publicly. That silence, he suggested, is part of the problem: internal unease is not translating into open debate about where the technology is heading. His decision to leave makes him one of a small but growing number of AI researchers willing to walk away from prominent positions over ethical objections.
The departure highlights a tension inside the AI industry, where researchers often work on foundational improvements that have broad and unpredictable effects. Chip design tools may seem distant from the headline-grabbing capabilities of chatbots and image generators, but they are part of the infrastructure that makes those systems possible. Faster, cheaper chips mean larger models can be trained more quickly, which in turn accelerates the cycle of capability gains.
Superintelligent AI — a hypothetical system that surpasses human intelligence across most domains — remains a subject of intense debate. Some researchers argue it could arrive within decades and pose existential risks if not carefully controlled. Others say such warnings are speculative and distract from more immediate harms, such as bias, misinformation, and labor disruption. O'Callahan's statement places him firmly in the camp that sees near-term superintelligence as a danger rather than a goal.
His resignation also raises questions about the responsibility of individual engineers and scientists within large technology companies. Even narrow technical contributions can have far-reaching consequences, and employees may find it difficult to predict how their work will be used. O'Callahan's conclusion — that he could no longer justify his own role — reflects a broader ethical reckoning that some in the field are beginning to confront.
Google DeepMind has not publicly commented on the resignation. The company, like other leading AI labs, has emphasized the importance of safety research and responsible development. But critics say voluntary principles are not enough when competitive pressures push companies to release powerful systems quickly.
For now, O'Callahan's exit adds another voice to the debate over whether the pursuit of advanced AI should be slowed, regulated, or redirected. His warning is unlikely to halt the industry's momentum, but it may encourage other researchers to weigh the consequences of their work more openly — and to ask whether silence is still an option.
5
