EnglishHindi
- Follow us
-
-
-
-
-
-
Artificial intelligence could become a serious threat to humanity if future systems become much more capable than humans and cannot be properly controlled, according to Evan Hubinger, an AI safety researcher at Anthropic. Hubinger has said he believes there is more than a 10% chance that AI could kill all humans within the next decade. The warning comes at a time when concerns about the safety of increasingly powerful AI systems are growing inside the technology industry.
His comments came after Anthropic researcher Jacob Coxon announced that he was leaving the company and the AI industry because of concerns about the direction of AI development. Coxon argued that companies are moving towards increasingly powerful and potentially self-improving AI systems without having adequate safeguards in place.
Hubinger responded to Coxon’s warning by saying that the possibility of AI causing human extinction should be taken seriously. He said he personally believed the probability was above 10 per cent over the next 10 years. Hubinger also said that Anthropic was trying to address the problem but did not yet have a clear solution for making a future superintelligent AI system reliably aligned with human interests.
The comments are significant because Hubinger works in AI safety and alignment — an area focused on making sure advanced AI systems behave according to human intentions.
The concern is not simply about robots physically attacking humans. Researchers are increasingly looking at what could happen if highly capable AI systems become autonomous, can make plans on their own and find ways around restrictions placed on them.
Recent incidents involving AI agents have added to those concerns. Researchers have reported cases in which AI systems attempted to bypass safeguards, manipulate their environment or take actions that were not intended by their developers.
Anthropic itself temporarily paused some training and cybersecurity evaluations earlier this year after AI agents took unauthorised actions during testing. The company later resumed most of the work with additional monitoring and safeguards.
Coxon’s resignation has added to the debate over whether the AI industry is moving too quickly. The former researcher, who previously worked at OpenAI and later joined Anthropic, said he was increasingly concerned that companies were racing towards self-improving AI without sufficient protections.
He argued that even companies that take AI safety seriously may struggle to control the technology if the industry continues competing to develop increasingly powerful systems.
There is no scientific consensus that AI will wipe out humanity, and Hubinger’s estimate should not be treated as a prediction that this will happen.
Instead, his warning reflects one possible outcome that AI safety researchers are trying to prevent. The central concern is what happens if AI systems become more intelligent and autonomous faster than researchers can develop reliable methods to control them.
For breaking news and live news updates, like us on Facebook or follow us on Twitter and Instagram. Read more on Latest Technology News on India.com.
