What is the story about?
Google DeepMind chief executive Demis Hassabis has renewed his warning that artificial general intelligence (AGI) could arrive within the next few years, while urging governments and the AI industry to move quickly on regulation before increasingly capable systems outpace existing safeguards.
In a new personal blog, the Nobel Prize-winning AI researcher argues that the world is approaching a defining moment in the development of artificial intelligence. Although he maintains that AGI is only "a few short years away", the prediction remains speculative and continues to divide the AI research community, with no consensus or evidence that human-level AI is imminent.
Hassabis defines AGI as "a system that exhibits all the cognitive capabilities the brain has." The timeline is consistent with comments he made in June, when he suggested AGI could emerge within three to four years.
Rather than focusing solely on AI's rapid progress, Hassabis argues that governments and technology companies are paying too much attention to the competitive race to build more powerful models and too little to the risks that accompany them.
He warns that frontier AI systems are already creating cybersecurity concerns and believes future generations of AI could introduce even greater challenges, including risks linked to biological and nuclear security.
"Urgent action is needed to address risks that might arise as we get closer to AGI. We've already seen the challenges frontier models pose for cybersecurity, and other threats including nuclear and bio risks may soon emerge as capabilities continue to advance," he wrote.
Hassabis also argues that researchers will eventually need robust mechanisms to retain control over increasingly autonomous AI systems, particularly if future models become capable of improving themselves with limited human involvement.
These concerns are shared by several leading AI companies. Anthropic, the developer of Claude, has repeatedly warned about the risks associated with advanced AI systems while acknowledging that researchers still do not fully understand how today's frontier models arrive at many of their decisions. Anthropic CEO Dario Amodei has similarly advocated stronger safeguards and regulatory oversight.
Despite the uncertainties surrounding AI development, Hassabis says he remains confident that the technical challenges can be addressed—but only if governments and the industry act before AGI becomes a reality.
As part of his proposal, Hassabis has called for the creation of a dedicated AI oversight body modelled on the Financial Industry Regulatory Authority (FINRA), the private organisation that regulates US brokerage firms under federal supervision.
The regulator, he suggests, would bring together leading technical experts, receive funding primarily from the AI industry and establish safety standards that frontier AI models must satisfy before they can be deployed.
Initially, developers would voluntarily submit their most advanced models for evaluation up to 30 days before release. Once testing procedures mature and prove reliable, participation would eventually become mandatory.
"Nobody in the world knows for sure what is going to happen from here, and even the experts disagree," Hassabis wrote. "When there is a large degree of uncertainty and the stakes are this high, proceeding with cautious optimism is the sensible and correct strategy."
He believes the United States is well positioned to launch such a framework because of its technological leadership, with the longer-term goal of establishing internationally accepted standards for frontier AI development.
While much of Hassabis's message centres on reducing risk, he also presents an optimistic vision of what AGI could eventually achieve if developed responsibly.
He argues that the technology has the potential to accelerate scientific discovery, transform healthcare through faster drug development, enable breakthroughs in clean energy and advanced materials, and help address some of humanity's most pressing challenges. In his view, AGI's long-term impact could surpass that of previous technological revolutions.
Reflecting on decades spent pursuing the technology, Hassabis writes that AGI should not be viewed as simply another computing milestone.
"It is much more akin to the discovery of electricity or fire. If you stop to think about it, we've essentially found a way to make sand think. It's miraculous."
At the same time, he acknowledges that the future remains uncertain and that experts continue to disagree over both AGI's timeline and its eventual capabilities.
"But the future is not yet written, we must use this precious window before AGI arrives to shape this technology for the benefit of all humanity," Hassabis wrote, arguing that decisions made over the coming years will play a critical role in determining how AI shapes the next phase of human civilisation.
In a new personal blog, the Nobel Prize-winning AI researcher argues that the world is approaching a defining moment in the development of artificial intelligence. Although he maintains that AGI is only "a few short years away", the prediction remains speculative and continues to divide the AI research community, with no consensus or evidence that human-level AI is imminent.
Hassabis defines AGI as "a system that exhibits all the cognitive capabilities the brain has." The timeline is consistent with comments he made in June, when he suggested AGI could emerge within three to four years.
Safety concerns take centre stage
Rather than focusing solely on AI's rapid progress, Hassabis argues that governments and technology companies are paying too much attention to the competitive race to build more powerful models and too little to the risks that accompany them.
He warns that frontier AI systems are already creating cybersecurity concerns and believes future generations of AI could introduce even greater challenges, including risks linked to biological and nuclear security.
"Urgent action is needed to address risks that might arise as we get closer to AGI. We've already seen the challenges frontier models pose for cybersecurity, and other threats including nuclear and bio risks may soon emerge as capabilities continue to advance," he wrote.
Hassabis also argues that researchers will eventually need robust mechanisms to retain control over increasingly autonomous AI systems, particularly if future models become capable of improving themselves with limited human involvement.
These concerns are shared by several leading AI companies. Anthropic, the developer of Claude, has repeatedly warned about the risks associated with advanced AI systems while acknowledging that researchers still do not fully understand how today's frontier models arrive at many of their decisions. Anthropic CEO Dario Amodei has similarly advocated stronger safeguards and regulatory oversight.
Despite the uncertainties surrounding AI development, Hassabis says he remains confident that the technical challenges can be addressed—but only if governments and the industry act before AGI becomes a reality.
Proposal for a US-led global AI regulator
As part of his proposal, Hassabis has called for the creation of a dedicated AI oversight body modelled on the Financial Industry Regulatory Authority (FINRA), the private organisation that regulates US brokerage firms under federal supervision.
The regulator, he suggests, would bring together leading technical experts, receive funding primarily from the AI industry and establish safety standards that frontier AI models must satisfy before they can be deployed.
Initially, developers would voluntarily submit their most advanced models for evaluation up to 30 days before release. Once testing procedures mature and prove reliable, participation would eventually become mandatory.
"Nobody in the world knows for sure what is going to happen from here, and even the experts disagree," Hassabis wrote. "When there is a large degree of uncertainty and the stakes are this high, proceeding with cautious optimism is the sensible and correct strategy."
He believes the United States is well positioned to launch such a framework because of its technological leadership, with the longer-term goal of establishing internationally accepted standards for frontier AI development.
Optimism tempered by uncertainty
While much of Hassabis's message centres on reducing risk, he also presents an optimistic vision of what AGI could eventually achieve if developed responsibly.
He argues that the technology has the potential to accelerate scientific discovery, transform healthcare through faster drug development, enable breakthroughs in clean energy and advanced materials, and help address some of humanity's most pressing challenges. In his view, AGI's long-term impact could surpass that of previous technological revolutions.
Reflecting on decades spent pursuing the technology, Hassabis writes that AGI should not be viewed as simply another computing milestone.
"It is much more akin to the discovery of electricity or fire. If you stop to think about it, we've essentially found a way to make sand think. It's miraculous."
At the same time, he acknowledges that the future remains uncertain and that experts continue to disagree over both AGI's timeline and its eventual capabilities.
"But the future is not yet written, we must use this precious window before AGI arrives to shape this technology for the benefit of all humanity," Hassabis wrote, arguing that decisions made over the coming years will play a critical role in determining how AI shapes the next phase of human civilisation.
