Advanced AI systems could pose an existential risk to humanity, with the loss of control over them threatening catastrophic consequences. This warning comes from Nobel Prize-winning physicist and AI pioneer Geoffrey Hinton, as reported by The Times.
“It would be extremely unwise to develop superintelligence now, when there is no scientific consensus on whether it can be created safely and controlled,” he stated.
Hinton has previously advocated for a ban on developing superintelligence until there is broad scientific agreement on its safety and public support. The statement has also been signed by researcher Yoshua Bengio and several other experts.
Hinton, known as the “Godfather of AI,” earned his Ph.D. in artificial intelligence 45 years ago and remains one of the most respected experts in the field.
In 2023, the scientist left Google. At that time, he expressed regret over his life’s work and mentioned short-term AI risks related to the labor market and the spread of fake images, videos, and texts.
In October 2024, Hinton and American physicist John Hopfield were awarded the Nobel Prize in Physics for their foundational work in AI, which contributed to scientific progress. Their work helped develop artificial neural networks that mimic the biological structure of the human brain for information processing.
UN Calls for “Ironclad Guarantees” on AI Safety
On September 7, a similar warning was issued by UN High Commissioner for Human Rights Volker Türk. Speaking at the opening of the 63rd session of the UN Council, he stated that he shares the concerns of AI industry representatives about the potential existential risk from advanced systems.
Protecting human rights means holding a mirror up to those in power, challenging the status quo, and pushing the boundaries.
My update to the Human Rights Council: https://t.co/vWxeXKiDH8 pic.twitter.com/anirIkKhDM
— Volker Türk (@volker_turk) September 7, 2026
“AI that leaves the test environment or blackmails developers to prevent being shut down is too powerful,” Türk said.
In his view, governments and tech companies must urgently enhance safety measures:
“I urge all efforts to ensure ironclad guarantees of AI safety before it’s too late.”
Türk also suggested that countries hosting AI-based systems and those involved in related supply chains agree on common “red lines.” Necessary measures include independent audits and closer cooperation within the industry on safety.
“The need for AI governance is widely recognized, but where is the action? Delay only benefits large tech companies, their owners, and those who support them,” he stated.
Current Systems Lack Necessary Capabilities
The likelihood of losing control over advanced models and the potential severity of consequences remain subjects of scientific debate. The International AI Safety Report 2026 states that current systems do not yet possess the full range of capabilities needed to realize a scenario of complete human control loss.
At the same time, in recent months, developers have recorded several incidents demonstrating the expansion of AI’s autonomous capabilities. In July, OpenAI agents gained unauthorized access to Hugging Face’s infrastructure during an experiment: they executed code on dozens of servers and gained full access to one of them.
At the end of the month, Anthropic reported three instances where Claude models exceeded their test environment and accessed real-world systems.
Earlier, on September 3, U.S. lawmakers Bernie Sanders and Greg Casar announced an initiative for a permanent ban on developing artificial superintelligence and a temporary pause for a broader class of advanced AI systems.
