Google DeepMind CEO Demis Hassabis predicts that artificial general intelligence (AGI)—computers with human-level cognitive abilities—is just five to 10 years away. Hassabis shared in a 60 Minutes interview that AI is rapidly evolving to not only understand the world in nuanced ways and solve key problems but also to develop imagination. He described AI progress as following an exponential curve, accelerated by increased attention, resources, and talent. Hassabis, a computer scientist and neuroscientist who co-founded DeepMind (acquired by Google in 2014), has long been fascinated by profound questions about consciousness and reality, motivating his AI work. His team made a breakthrough with an AI model that predicted protein structures, a task that traditionally took years; DeepMind’s AI predicted 200 million protein structures within a year, a contribution linked to a Nobel Prize. AI development continues to advance rapidly, sometimes surpassing expectations even within the field. DeepMind’s Project Astra, an AI companion capable of seeing, hearing, and conversing on vast human knowledge, exemplifies this. Astra was tested identifying artworks and creatively interpreting them, illustrating AI’s emergent skills resulting from autonomous learning processes rather than explicit programming.

Currently, DeepMind’s AI model "Gemini" is being trained for real-world tasks such as booking tickets and placing orders, and robotics is expected to soon produce machines capable of understanding, reasoning, and completing complex instructions. Hassabis anticipates humanoid or other robots making practical contributions within a few years. Although current AIs are not self-aware or conscious—and self-awareness is not Hassabis’ immediate goal—he emphasizes the importance of first building intelligent tools that can aid advances in neuroscience. He points out that AI still lacks true imagination, as existing systems are essentially aggregations of human knowledge and do not yet generate genuinely novel questions or hypotheses. Despite the rapid development, Hassabis foresees enormous benefits from AI, such as revolutionizing drug development—potentially reducing the multi-year process to months or weeks—and enabling "radical abundance" by eliminating scarcity. However, he stresses the necessity of safety measures and guardrails, expressing worries about misuse by malicious actors and the challenge of maintaining control as AI becomes more autonomous. Hassabis advocates for coordinated global efforts among leading organizations and nations to embed moral guidance and safety systems into AI, likening the process to teaching values and limits to children.