AI Researchers Sound the Alarm on the Risks of Uncontrolled Artificial Intelligence
When the very people building artificial intelligence begin to walk away in fear, it signals a critical moment for the industry and society at large. Jacob Coxon, a prominent AI researcher who recently moved from OpenAI to Anthropic, has made headlines by quitting the AI field altogether. His departure stems from deep concerns that the rapid development of AI systems is outpacing our ability to control them safely.
Coxon expressed his fears in an interview with The Wall Street Journal, warning that “we’re on track for a lot of the most aggressive of these scenarios.” He cautioned that by the end of next year, AI development could already be “out of control.” This stark prediction highlights the urgent need for a more measured approach to AI advancement.
From OpenAI to Anthropic: A Search for Safer AI Development
Jacob Coxon’s transition from OpenAI to Anthropic earlier this year was driven by Anthropic’s reputation for prioritizing AI safety. Anthropic has gained recognition for its attempts to develop safer AI systems, emphasizing ethical considerations and risk mitigation. However, Coxon’s latest stance reveals his growing skepticism even about Anthropic’s capacity to manage the risks inherent in powerful AI.
He now argues that without meaningful government intervention or an industrywide slowdown, it may be impossible to responsibly pursue AI capable of outperforming humans across a broad range of tasks. This perspective resonates with calls from many experts for stronger regulatory frameworks and collaborative efforts to ensure AI development does not spiral beyond human control.
Industry Voices Echo the Warning
Supporting Coxon’s concerns, Evan Hubinger, Anthropic’s Alignment Science Lead, made a startling statement via Twitter that has attracted national attention. He wrote, “We really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade.” Hubinger further admitted that while Anthropic is trying its best, the company “does not yet have a plan to solve alignment for superintelligence and are not clearly on track to.”
Jacob is correct here—we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade. I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to.
— Evan Hubinger (@EvanHub) September 9, 2026
Their warnings are not isolated. Across the industry, researchers and executives at leading AI companies—including OpenAI, Anthropic, and xAI—have departed amid growing concerns about safety, ethics, and the unchecked speed of AI progress. According to multiple reports, these insiders emphasize the need for urgent safeguards to prevent potentially catastrophic outcomes.
The Urgent Need for Ethical Oversight and Regulation
The rapid evolution of AI technologies presents unprecedented challenges and risks. Experts like Coxon and Hubinger underscore that without robust alignment strategies and comprehensive regulatory measures, AI systems could evolve beyond human control, with potentially devastating consequences.
Governments, industry leaders, and the global community must collaborate to establish frameworks that prioritize safety and ethics in AI development. This includes slowing down the deployment of advanced AI models until reliable alignment methods are validated, increasing transparency, and fostering accountability among AI creators.
As the stakes grow higher, the voices of those who understand AI’s inner workings—like Coxon and Hubinger—offer critical insight into the potential dangers ahead. Their candid warnings serve as a call to action for policymakers, technologists, and society to tread carefully in this transformative era.
For more detailed coverage on this developing story, visit Here.
