DeepMind CEO Says AGI Will Be Bigger Than Electricity Or Fire

For the second time this year, Demis Hassabis has predicted that artificial intelligence will arrive by the end of the decade. However, this time, he said that it will not be another technical skill – it will fight against the presence of electricity or fire.

In a blog post published on Tuesday on X, the CEO of Google DeepMind said that AGI is “probably only a few years away,” explaining that it is a technology that can change the development of humanity.

“When we look back at this time in the next few decades, I think we will realize that we were standing on the precipice of a single species – not just the beginning of a new generation of people.”

According to Hassabis, AGIthe fact that computers can understand, learn, and perform various tasks as well as or better than humans, should not be compared to advances such as the Internet or mobile computers because the consequences can be very large.

He wrote: “It is very similar to the presence of electricity or fire. “When you stop and think, we have found a way to make sand think.

Despite that hope, Hassabis warned that AI technology is advancing faster than people’s ability to understand and manage risks, pointing to cyber security threats that already exist with today’s models, adding that future systems could cause natural, nuclear, and other national security risks.

As AI becomes more stable and capable of self-correction, he said, strong technological safeguards will be needed to ensure that humans remain in control.

“In the near future, we will need strong security to keep up with these increasingly complex, automated systems—and to deal with unknown challenges that will only become clearer over time.”

The news comes as AI leaders mark the years since ChatGPT was launched in 2022 and warn that AGI may arrive sooner than expected. In January 2026, Anthropic CEO Dario Amodei said human level AI may appear within one to five years and warn governments of slowing development. Then, in June, Hassabis he predicted AGI will arrive by 2030 and warns people that they “don’t take long to prepare.”

To address these concerns, Hassabis decided to create a US Frontier AI Standards Body modeled after the Financial Industry Regulatory Authority, or FINRA, the trade body that oversees US financial institutions. The public-private partnership will be supported primarily by AI companies and supported by independent technical experts and open source representatives to evaluate frontier AI models.

“The progress we’re seeing in AI requires a new way to test AI modeling skills that is dynamic, flexible, and robust,” he wrote. “The US is in a good position, both economically and technically, to play a role in creating such a system.”

The proposal follows a similar call made by prominent members of the industry to establish oversight of advanced AI.

In May 2023, at a meeting of the US Senate Committee on the Judiciary, the CEO of OpenAI Sam Altman. to be invited for the federal agency to issue licenses for powerful AI systems and require independent safety reviews. Most recently, last month, President Donald Trump signed it Executive Order creating a voluntary process for evaluating advanced AI models before they are released. That same month, Anthropic CEO Dario Amodei he warned that AI is becoming more powerful and safety regulations such as the Federal Aviation Administration (FAA) are needed.

Despite the push to regulate AI development, Hassabis said the world has a limited window to establish common standards before AGI arrives.

“The future is unwritten, we must use this precious window before AGI arrives to develop this technology for the benefit of all,” he wrote. “What we do together now will shape the future of development. By managing AGI around the world, we can enter a new era of scientific discovery and progress, and bring about a better future for human development.”