Human-level Artificial Intelligence  could arrive within a decade

Artificial general intelligence (AGI) — machines with human-like cognitive abilities — may be just five to ten years away, according to Demis Hassabis, CEO of Google DeepMind and Nobel Prize-winning AI researcher.

In a recent interview on CBS News’ 60 Minutes, Hassabis offered both a bold vision for AI’s future and a grounded assessment of its current limits. While the path toward AGI appears increasingly within reach, he clarified that today’s AI systems lack consciousness and imagination.

“They’re still essentially the average of all human knowledge they’ve been trained on,” he said. “They can’t yet ask truly novel questions or generate hypotheses that have never been thought of.”

Hassabis advised a cautious and strategic approach to AI advancement: “We should focus on building intelligent tools to accelerate neuroscience before tackling the question of self-aware machines.”

The DeepMind CEO is optimistic about AI’s near-term impact, especially in healthcare. He believes AI could revolutionize drug development, drastically cutting costs and timelines. “It currently takes about 10 years and billions of dollars to create a single drug. AI could potentially reduce that to months or even weeks,” he said.

Looking further ahead, Hassabis pointed to robotics as AI’s next frontier. “In the next few years, we’ll likely see breakthroughs — robots, possibly humanoid, that can perform genuinely useful tasks by understanding the world around them.”

Hassabis, a former chess prodigy turned neuroscientist, co-founded DeepMind in 2010. In 2024, he and colleague John Jumper received the Nobel Prize in Chemistry for their development of AlphaFold, an AI that accurately predicts protein structures — a milestone with transformative implications for science and medicine.