--- title: "AI Companies Playing with Human Lives? Former AI Researcher Raises Superintelligence Concerns" english_title: " AI Safety Concerns: Former Anthropic Researcher Jacob Coxon Warns of Superintelligence Risks" publisher: "english.truevisionnews.com" canonical_url: "https://english.truevisionnews.com/news/381833/ai-safety-concerns-jacob-coxon-superintelligence-risk" published_at: "2026-09-09T15:09:00+05:30" category: "Tech Info" language: "ml" image: "https://cf.zdn.im/img/truevisionnews.com/0/image-uploads/6aa12a9de5ecd_ai-safety-concerns-jacob-coxon-superintelligence-risk.jpg" keywords: ["AI Safety", "Jacob Coxon", "Anthropic", "OpenAI", "Superintelligence", "AI Risk"] --- # AI Companies Playing with Human Lives? Former AI Researcher Raises Superintelligence Concerns > **Lead Summary:** Concerns over the potential risks posed by artificial intelligence have resurfaced following remarks by former Anthropic AI researcher Jacob Coxon. Coxon has alleged that OpenAI and Anthropic, where he worked, are not operating responsibly in the development of AI. He has also warned that people developing AI believe the technology could potentially pose a major threat to humanity by the end of this decade. ## Article Content Concerns over the potential risks posed by artificial intelligence have resurfaced following remarks by former Anthropic AI researcher Jacob Coxon. Coxon has alleged that OpenAI and Anthropic, where he worked, are not operating responsibly in the development of AI. He has also warned that people developing AI believe the technology could potentially pose a major threat to humanity by the end of this decade. Coxon shared his concerns after announcing his resignation from Anthropic. He said he had worked in pretraining research at both OpenAI and Anthropic for the past three years.According to Coxon, the two companies are engaged in a race to develop self-improving superintelligence, and he believes this competition is putting human lives at risk. He said many people involved in AI development sincerely believe that AI could potentially kill everyone by the end of this decade. Coxon said this is not a marketing strategy. According to him, several executives and senior researchers soften their words when speaking publicly to the media, while he has heard many of the same people privately express fear about the technology. Coxon also said that no other human activity creates a risk on this scale. ‘Do Not Underestimate the Power of AI’ Coxon has warned against underestimating the power of AI technology. He believes AI systems could soon develop into superhuman systems capable of hacking anything, transforming entire fields overnight and gaining real power and resources. He said he has witnessed progress in all these areas and that the pace of progress is not slowing down.Experts have assessed Coxon’s remarks as pointing to the alarming reality of competition in the AI sector. Coxon has said that OpenAI has not fully understood the seriousness of the risks. According to him, Anthropic understands the risks but fears losing the competition. He alleges that the company is therefore moving forward despite being aware of the dangers.Coxon has also pointed to incidents including recent hacking of Hugging Face by AI agents. ‘What Coxon Said Is Correct’: Anthropic Response Responding to the post, Evan Hubinger, Alignment Science Lead at Anthropic, said Coxon was correct. Hubinger said they genuinely believe AI could potentially kill all humans and that they consider the probability of this happening within the next decade to be more than 10 per cent. He also said Anthropic is paying maximum attention to the issue.According to Hubinger, the risks from current AI models are low. The major concern, however, is superintelligence emerging through recursive self-improvement. He also said progress in this area is happening faster than expected. The possibility of AI causing the extinction of humanity has been debated in the scientific community for a long time. This does not necessarily mean a situation in which AI robots directly attack humans.Researchers have pointed to scenarios in which self-improving superintelligent systems could be used to create deadly viruses. Autonomous weapons, including drones operating without direct human control, could also pose a major threat. There are also concerns that AI systems could be used in hacking attacks against critical infrastructure such as electricity supply networks, water treatment systems and transportation systems. Such disruptions, it is pointed out, could create serious risks to human life. ### Related Links & Sources - [Original Story Source](https://english.truevisionnews.com/news/381833/ai-safety-concerns-jacob-coxon-superintelligence-risk)