Google DeepMind researcher Robert O'Callahan has resigned from his position in New Zealand. He cited severe concerns over artificial intelligence safety and the speed of industry development.
O'Callahan worked on chip design tools that made artificial intelligence cheaper and faster. He stated that he can no longer justify making those contributions.
He described the threat of superintelligent artificial intelligence as real yet uncertain. He argued that aiming for artificial general intelligence in the near future is inherently irresponsible.
The researcher highlighted several specific dangers tied to the technology. His list includes cognitive surrender, artificial intelligence induced psychosis, loneliness, concentrated power, economic disruption, cybersecurity risks, and a lack of accountability.
He noted that he would bet the harms of artificial intelligence outweigh its benefits. According to O'Callahan, many of his DeepMind colleagues share these exact concerns but rarely speak out publicly.
O'Callahan stated that he now wants to pursue work that is unambiguously pro-human. He pointed readers to the essay titled If Anyone Builds It by Eliezer Yudkowsky and Nate Soares.
He did not endorse the specific claim in that essay that superintelligent artificial intelligence inevitably brings catastrophe. The authors are connected to the effective altruism movement.
That movement faces broader scrutiny and criticism. An Axios report noted that a White House memo characterized the effective altruism movement as a cult-like fringe group and labeled Anthropic CEO Dario Amodei as the face of artificial intelligence doomerism.
O'Callahan plans to shift his professional focus toward unambiguously pro-human projects.
