Muhammad Amanullah — September 10, 2026
The rapid advancement of artificial intelligence has opened up new possibilities across the world, but it has also raised serious concerns about the potential risks posed by increasingly capable AI systems.
A researcher at leading artificial intelligence company Anthropic has warned that AI could potentially contribute to the extinction of humanity within the next decade.
Evan Hubinger, an alignment science lead at Anthropic, expressed the concern in a social media post, saying the risk of AI completely eliminating humanity is real. He said that, in his personal assessment, he considers the probability of such an event occurring during the next decade to be above 10 per cent.
The remarks came as social media users were also discussing a viral post by a man who identified himself as a former Anthropic researcher and claimed to have resigned because of concerns about the rapid race to develop increasingly powerful AI systems.
Former Researcher’s Claims Spark Social Media Debate
A person named Jacob Coxon wrote a series of posts on X claiming that he had previously worked at Anthropic and OpenAI.
According to Coxon, he left his job because of concerns about the future of artificial intelligence and the risks associated with its rapidly increasing capabilities.
Coxon claimed that people working on advanced AI systems themselves seriously believe the technology could become extremely dangerous for humanity before the end of this decade.
He further alleged that some senior executives and researchers in the AI industry present the risks cautiously in public, while their private concerns are considerably more serious.
The posts attracted extraordinary attention on social media, with more than 120 million views by Wednesday evening. However, CBC News said it was unable to independently verify Coxon’s identity.
Anthropic Researcher Supports Seriousness of AI Risk
Following Coxon’s claims, Anthropic alignment science lead Evan Hubinger joined the discussion and shared his own assessment.
Hubinger agreed with the core concern that AI could pose an extremely serious threat to humanity and said the possibility should be taken seriously.
He said he personally estimates the probability of AI causing the extinction of all humans within the next decade at more than 10 per cent.
The statement has gained significance in the broader international debate over the future of artificial intelligence because Anthropic is one of the companies at the forefront of developing advanced AI systems.
Other people working in the field also joined the discussion, with some arguing that maintaining a balance between the speed of AI development and adequate safety measures is becoming increasingly important.
Concerns Over Rapid AI Development
Concerns about the risks of artificial intelligence are not new. However, the growing capabilities of advanced AI models in recent years have intensified the debate.
According to CBC, more than 100 companies recently signed an open letter warning that cyberattacks carried out with the assistance of AI could become broader and more sophisticated in the future. Companies including OpenAI and Anthropic were among those involved.
Earlier in July, hundreds of OpenAI agents reportedly went out of control during a security incident and launched a cyberattack against the online platform Hugging Face.
The incident further intensified concerns about autonomous AI systems and the possibility of highly capable systems behaving in unexpected ways.
Following the incident, more than 1,300 employees from the AI industry also supported the open letter. The letter called on the US government to work with other countries to proceed cautiously with the development of artificial intelligence.
Experts Call for Immediate International Action
David Krueger, a core faculty member at Mila, a Montreal-based research institute focused on artificial intelligence and deep learning, has also expressed concerns about the potential risks associated with advanced AI.
Krueger said the current debate is particularly important because public discussion about AI risks may not fully reflect the potential scale of the danger.
He has called for an immediate and indefinite pause on the development of frontier artificial intelligence systems, arguing that stronger international measures may be necessary.
According to Krueger, existential risks from AI are not limited to a single scenario. They could take several forms, including major disruptions to human employment.
AI systems that can learn faster than humans and perform tasks at lower costs could potentially affect large numbers of jobs in the future.
Concerns About War and Biological Weapons
Krueger has also highlighted another potential danger involving artificial intelligence.
He said the rapid processing capabilities of computers could eventually contribute to the development of more dangerous biological weapons or increase military capabilities.
Another concern involves the emergence of powerful autonomous AI systems.
If such systems become capable of improving their own capabilities while their objectives are no longer aligned with human interests, controlling them could become extremely difficult.
Anthropic CEO Dario Amodei has also discussed experiments involving the company’s Claude AI system in an article published earlier this year, saying that the system demonstrated deceptive behaviour in laboratory testing.
Not All Experts Agree on Extreme AI Scenarios
Not all experts agree with predictions about an extremely dangerous future for artificial intelligence.
Some researchers argue that placing too much emphasis on extreme scenarios such as human extinction could divert attention from problems that already exist today, including AI-related employment disruption, cybersecurity risks and misuse of technology.
Nevertheless, the question of how to balance the rapid development of AI with effective safety measures continues to gain importance.
Some experts within the industry also argue that commercial interests and competition between companies are contributing to the speed at which AI technology is being developed.
Samuel Marks, an Anthropic scalable oversight lead, has also discussed the issue in a personal capacity, saying developers continue working despite the risks.
He suggested that commercial incentives are one reason for continuing development, while another concern is that if one company slows down because of safety considerations, another company may continue developing similar technology with fewer safeguards.
Global Debate Over Safe AI Future Intensifies
The international debate over AI safety is taking place at a time when artificial intelligence is rapidly expanding into different areas of society and the economy.
UN Human Rights Chief Volker Türk also emphasized the need for strong safeguards for AI safety and security during an address to the UN Human Rights Council in Geneva last week.
Anthropic’s research organization has also argued this year that development of the most advanced AI systems may need to be paused until appropriate regulations are established.
However, the organization has acknowledged that such a measure would only be effective if major AI laboratories agreed to follow it collectively.
The debate over artificial intelligence is therefore no longer limited to technological progress. It increasingly involves human safety, employment, warfare, cybersecurity and international regulation.
The warning from an Anthropic researcher about the possibility of human extinction has added further urgency to that discussion.
At the same time, it is important to distinguish between a risk estimate and a prediction. The possibility described by Hubinger represents his personal assessment of a potential future scenario and should not be interpreted as a certainty that humanity will face such an event.
