Evan Hubinger, a leading safety researcher at Anthropic, has sounded the alarm on social media, stating that he personally believes there is a greater than one‑in‑ten risk that advanced artificial intelligence could wipe out humanity within the next ten years. His remarks, posted on X and viewed more than ten million times, come amid revelations that the firm withheld its newest model from the United Kingdom’s AI Safety Institute, prompting fresh debate over how aggressively the technology is being pursued.
Researcher’s Stark Warning
In his post, Hubinger acknowledged that the danger posed by today’s models remains “low”, but he warned that the real threat lies in the prospect of AI soon gaining the ability to improve itself. He said, “we really do earnestly believe” AI poses a species‑ending risk to humans, and added, “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,”. The warning echoes a growing unease among AI specialists that the pace of development may be outstripping our capacity to keep the technology under control.
Withheld Model Raises Transparency Questions
The Financial Times reported that Anthropic chose not to submit its latest system to the UK’s AI Safety Institute, one of the world’s foremost bodies for evaluating AI risk. A Cabinet Office spokesperson declined to confirm whether the model had been withheld, noting only that the government “continues to collaborate closely with industry partners, including Anthropic, to make models safer”. The move has intensified scrutiny over how leading AI labs balance rapid innovation with safety oversight, especially as the UK positions itself as a global hub for AI governance.
Industry Voices React
Neil Lawrence, Professor of Machine Learning at the University of Cambridge, told BBC Radio 4’s Today Programme that the report was credible, suggesting that the United States’ competitive stance vis‑à‑vis China could be prompting a retreat from international cooperation. He remarked, “I suppose it’s unsurprising against a background where there’s a perception where the United States very much sees AI as a race between themselves and China and is moving more towards isolationist positions, that it might be that the administration is saying that they should reduce cooperation with some of their allies,”. Meanwhile, OpenAI’s chief scientist Jakub Pachocki urged “extreme caution” over AI’s progress, warning that further intervention may be needed to ensure “humans remain in control of the future”. An open letter signed by 1,300 employees across AI firms called on the US government to “support an international effort to develop the technical and governance tools needed to deliberately pace the frontier of automated AI development”.
Why it Matters
Hubinger’s warning underscores a pivotal moment for the tech industry: the line between breakthrough gadgets and existential threats is becoming increasingly blurred. If advanced AI can indeed evolve beyond human oversight, the stakes extend far from consumer gadgets to the very survival of our species, demanding urgent, coordinated action from developers, governments, and global safety institutions.
