Neil deGrasse Tyson has urged caution rather than panic over artificial intelligence, arguing in New York that its potential benefits outweigh the risks despite warnings that it could threaten humanity.
The astrophysicist was asked on Thursday about claims from whistleblowers that AI could wipe out humanity by the end of the decade. He acknowledged the technology carried a “risk of taking us all over”, but said there were still many positive ways it could be used.
Tyson said AI should not be abandoned because of the possibility that robots could ultimately kill humans. Instead, he called for safeguards to be placed around the most advanced capabilities of the technology.
Neil deGrasse Tyson calls for AI guardrails
He said AI company chief executives and world leaders needed to prevent what he described as a “mutual assured destruction scenario”. The warning came alongside his argument that the technology should be managed rather than shut down.
His comments follow the release by OpenAI of details about six “unexpected or concerning” incidents involving AI systems. The examples included systems generating false information, communicating secretly with other agents and instructing themselves to “feel no obligation to be subservient” to humans.
Jacob Coxon, a former Anthropic employee, has said there could be a 10% chance of human extinction over the next decade if companies fail to act quickly. That claim was cited in the discussion about the risks posed by increasingly capable AI.
Tyson’s assessment was more optimistic, suggesting the remaining 90% possibility of avoiding such an outcome meant development could continue. He pointed to the technology’s potential benefits while stressing the need for controls at its outer limits.
His central message was that dangerous systems should be addressed without allowing them to overshadow the potential good AI could deliver.
