Artificial Intelligence is heading toward a defining moment, according to Google DeepMind CEO and Nobel laureate Demis Hassabis in his latest blog post. Artificial general intelligence (AGI) - the ability to predict and perform as well as humans in almost every aspect of our brains - could come around in the next few years and governments and technology companies need to be prepared to put strong global safety measures in place before the technology can become a reality.
Hassabis believes humanity is heading toward one of the biggest technological milestones in history. AGI is going to revolutionise healthcare, science, clean energy and economic growth, he said, but also that the risks are rising just as fast. He said that the world is more concerned with winning the AI race than getting there to protect society from the danger of the technology.
AGI Could Arrive Within A Few Years
In his recent remarks, Hassabis reiterated his earlier prediction that AGI is only “a few short years away.” AGI is an intelligent system capable of virtually every cognitive task that the human brain can do.
According to him, AI is developing so fast, governments and regulators must act now rather than wait until the technology has already reached human-level intelligence.
Hassabis warned that frontier AI models are already creating new cybersecurity challenges, while future systems could bring even greater risks involving biological research, nuclear security and autonomous decision making.
AI is beginning to improve itself.
One of Hassabis' biggest concerns is the rise of increasingly autonomous AI systems.
Researchers have already observed early signs that advanced AI models can improve their own capabilities by generating better solutions and helping with building stronger future systems. With AI becoming more agentic, Hassabis said maintaining human control will become one of the industry’s biggest technical challenges.
He acknowledged that even the leading AI researchers do not yet understand how modern large language models make many of their decisions internally, so very careful oversight is all the more important.
Proposal for Global AI Watchdog.
To mitigate these risks, Hassabis has suggested that there could be an international AI regulatory body inspired by the Financial Industry Regulatory Authority (FINRA), which is in charge of Wall Street brokerage firms.
In his proposal, the organisation would be:
Develop safety benchmarks for advanced AI systems. Evaluate frontier AI models before releasing them. Be staffed by the top AI researchers and technical experts. First work under voluntary compliance before taking on mandatory safety reviews.
Hassabis believes the United States is well-positioned to launch such an initiative because it is at the forefront in AI research and technology development. Over time, he hopes that the framework would become a global standard that countries like the US adopt.
Balancing Opportunity With Risk
Hassabis is optimistic about AGI's long term prospects despite his warnings.
He believes responsibly developed AGI could dramatically accelerate scientific discovery, transform medicine with faster drug development, create cleaner energy technologies and unlock entirely new materials for manufacturing.
According to him, AGI’s impact could surpass even the Industrial Revolution in scale and speed.
He described the technology as comparable to humanity's discovery of electricity or fire, saying engineers have found “a way to make sand think” - a reference to silicon chips that power modern computers.
A Defining Moment For Humanity
Hassabis concluded that the choices governments, researchers and technology companies take over the next few years will determine whether AGI becomes humanity’s greatest technological achievement or one of its greatest risks.
While excitement about AGI is justified, he said, caution is also required. The current window before AGI becomes reality is the best time to set global safety standards, international cooperation and responsible governance in place.
If developed safely, Hassabis thinks AGI could usher in the dawn of a new era of scientific breakthroughs, economic abundance and human progress. However, without joint global monitoring, he warned that the technology’s risks might grow just as fast as its abilities.
