DeepMind’s CEO Says AGI Could Arrive Around 2030. The Scary Part Is How Unready We Are

Demis Hassabis says AI is entering a “species-level transition.”

Demis Hassabis does not think today’s chatbots are the endgame.

The Google DeepMind CEO says artificial general intelligence, or AGI, could arrive around 2030. Hassabis runs one of the world’s most important AI labs, and DeepMind has already shown how machine learning can change science, most famously with the Nobel Prize-winning AlphaFold, which solved the 50-year-old “protein-folding” problem. It took thousands of scientists decades to map the structure of 150,000 proteins while AlphaFold cataloged the predicted structures of over 200 million proteins — encompassing nearly every known protein on Earth.

So, Hassabis knows a thing or two about AI and when he speaks about it, I tend to listen carefully.

“I believe that we’re only a few years away from that, maybe 2030, plus or minus a year — which is astounding to think, really,” Hassabis said during an interview at the Stanford Graduate School of Business. “I think that will be such an enormous transformative technology,” he added. “It’s going to effectively be a new human era.”

AGI is supposed to match or exceed human cognitive abilities across any intellectual task, from coding and scientific research to creative arts. Unlike current “narrow” AI, which is trained for specific functions like translation or chess, AGI would possess autonomous, generalized adaptability — meaning it could learn, synthesize, and apply entirely new skills without human reprogramming. An AGI would be able to autonomously solve some of the world’s most intractable problems and be able to improve itself.

This is a true paradigm shift because it moves humanity into an era of virtually infinite supply of cognitive labor that can accelerate scientific discovery, automate complex industries, and reshape global economics overnight. No one knows what the world would look like when such a god-like technology surfaces, but one can only expect nothing will ever be the same.

If Hassabis is right, society has only a few years to prepare for a technology that could reshape work, security, and daily life.

Not Just Another Chatbot

During his recent fireside chat at Stanford Graduate School of Business, Hassabis described AI as entering a “species-level transition.” He also said humanity has “little margin for error” over the next decade.

×

Get smarter every day...

Stay ahead with ZME Science and subscribe.

Daily Newsletter
The science you need to know, every weekday.

Weekly Newsletter
A week in science, all in one place. Sends every Sunday.
No spam, ever. Unsubscribe anytime. Review our Privacy Policy.

Thank you! One more thing...

Please check your inbox and confirm your subscription.

He argues that the next major step will come from AI agents. These are systems that do not simply respond to a prompt in a chat window. They carry out chains of tasks in the real world.

In his examples, an AI agent could book a weeklong vacation across 20 websites. A more powerful version could help run a drug-development research program.

RelatedPosts

Google’s Neural Machine can translate nearly as well as a human
Google Trends data can predict football player’s market value
You can now use Google Maps to explore other moons and planets
Scientists Engineered a Tobacco Relative to Produce Five Psychedelics at Once

In fact, Hassabis sees the most upside in research. He described AGI as the “ultimate tool for science” and summed up DeepMind’s mission in its winning formula: “Step one: solve intelligence. Step two: use it to solve everything else.”

The Industry Does Not Agree on the Timeline

Hassabis is not the only AI leader making bold predictions. OpenAI CEO Sam Altman wrote that “humanity is close to building digital superintelligence.” Anthropic CEO Dario Amodei predicted in 2024 that AGI could arrive as soon as 2026-2027. However, tech leaders have an incentive to embellish the pace of AI progress as this can help them secure more funding or grow their stocks.

Yann LeCun, Meta’s former chief AI scientist and one of the pioneers of modern machine learning, is more dismissive. The “concept of general intelligence is complete BS,” he said, according to Fast Company. LeCun argues that today’s large language models are unlikely to reach human-level intelligence or produce the kind of high-value work that AGI boosters imagine.

Even inside DeepMind, the forecasts differ. Shane Legg, DeepMind’s chief AGI scientist, has predicted a 50 percent chance that “minimal AGI” arrives in 2028. “Minimal AGI” means AI that can complete some cognitive tasks humans can do, but it can’t be better at anything than any human.

Hassabis tends to agree that some in the industry are “way too certain” about their predictions, he told the audience gathered at Stanford.

Still, he keeps returning to the same warning: preparation cannot wait.

“I still think there’s a lot more work, and it’s just the beginning,” Hassabis said. “I think society needs to hear that, because we don’t have long to prepare for what that means, and it’s going to be enormously profound.”

The Upside Is Huge. So Are the Risks.

The optimistic case for AGI is easy to understand. It could accelerate medical research, help design new materials, improve climate models, push fusion energy forward, and make scientific discovery faster and cheaper. An AGI used for good could lead to a “post-scarcity world,” a long-running futurist idea in which technology makes goods and services so abundant that everyone will be content from a material standpoint.

“Humans should always maintain their sense of meaning and what they decide to focus their lives on,” Hassabis said. “We shouldn’t become this kind of passive recipient of the technology.”

However, in the wrong hands, a frontier AGI could help bad actors design pathogens, automate cyberattacks, scale disinformation to frightening levels, or even control every human being with an all-seeing eye. That is why Hassabis supports “smart, targeted” regulation, including independent evaluations of model capabilities.

Even before an AI overlord runs the world, the labor question is already a pressing issue. AI writes code, drafts documents, and automates many office tasks. AI-driven efficiencies have contributed to more than 150,000 job cuts globally in the first half of 2026 alone. And that vastly outpaces the previous year’s totals. The tech sector has felt the immediate brunt of this transition. For instance, The Guardian reports that Amazon shed 30,000 employees over a six-month window as part of an aggressive operational realignment. Meanwhile, payment giant Block downsized its staff by 40% in February following executive directives to lean into AI-fueled productivity. Social media powerhouse Meta slashed 10% of its workforce to aggressively fund its multi-billion-dollar AI infrastructure, a trend mirrored by smaller tech players like Cloudflare and ed-tech firm Chegg.

Altman has warned that AI could erase large categories of jobs. Amodei has said half of entry-level white-collar work could disappear within five years, according to Business Insider.

For his own part, Hassabis thinks there will still be room for humans even in a job market dominated by powerful AI systems. He argues that people with “taste, design sensibility, original thinking” and the ability “to synthesize different subjects together” will be in a strong position over the next five years.

“I think amazing new things are going to be created,” Hassabis said. “I’ve got a lot of faith in human ingenuity, where we’re sort of instantly adaptable.”

Then he added: “We’re general intelligences ourselves, don’t forget. Look at what we built around us — it’s incredible — with our hunter-gatherer brains. Why would we stop here?”