DeepMind CEO: AGI, AI’s Limits.
DeepMind Chief Predicts Human-Level AI Within 10 Years
According to Demis Hassabis, CEO of Google DeepMind, artificial general intelligence (AGI)—AI capable of thinking and reasoning like a human—could emerge within the next five to ten years.
While the progress is promising, Hassabis emphasized that current AI systems are still far from conscious thought. He pointed out that today’s AI lacks genuine creativity and cannot yet demonstrate imagination.
In an interview aired on 60 Minutes by CBS News on April 22, Hassabis stated that these systems essentially mirror the vast pool of existing human knowledge. “They’re not yet capable of asking entirely new questions or forming original ideas that haven’t been previously considered,” he remarked.
Demis Hassabis believes that with the help of AI, the end of many diseases could be within reach over the next decade. AI-driven advancements, he explained, are accelerating the pace of drug discovery. “Typically, it takes about 10 years and billions of dollars to develop a single drug. With AI, we could potentially cut that down to months or even weeks,” he noted.
He also highlighted that the future of AI will be deeply integrated with robotics. In the coming years, Hassabis expects to see significant progress, including robots—possibly humanoid—that can perform practical, real-world tasks. “We’re approaching a breakthrough where robots will start to truly interact with and understand their surroundings,” he said.
A renowned computer scientist with a background in neuroscience, Hassabis studied at Cambridge, MIT, and Harvard. He co-founded DeepMind, an AI research company that was later acquired by Google in 2014. Alongside fellow researcher John Jumper, he developed AlphaFold, a groundbreaking AI model that accurately predicts protein structures. Their work earned them the 2024 Nobel Prize in Chemistry.
In the same year he received the Nobel Prize, Demis Hassabis was also knighted by King Charles of the United Kingdom.
Commenting on the rapid evolution of the AI sector, Hassabis noted, “We seem to be on an exponential growth curve. The success of recent breakthroughs has brought in more talent, funding, and attention, further accelerating this momentum.
Despite these remarkable advancements, Hassabis emphasized that AI remains something of a mystery.
Also read: Infinix Note 50s 5G Goes Official.
“We can predict certain capabilities based on the design of these systems, but ultimately, the learning process is similar to how humans learn. We don’t program every detail—they absorb patterns from data during training. That’s why unexpected behaviors or skills can sometimes emerge,” he explained.
He also underscored the importance of setting boundaries for AI, comparing its development to raising a child. “It’s essential to instill a value system, provide direction, and set safety limits—just like you would guide a young mind,” he said.
