Watch: Demis Hassabis, CEO of Google DeepMind and AI Nobel Prize winner Demis Hassabis, on CBS’ “60 Minutes”

A CBS weekly in-depth television news program Last night’s 60 Minutes ( Also shared on YouTube), this interview gave an inside look at Google DeepMind’s vision and the vision of its cofounder and Nobel Prize winning CEO, legendary AI researcher Demis Hassabis. The interview traced DeepMind’s rapid progress in AI and its ambition to reach artificial general intelligence (AGI), a machine intelligence that has human-like versatility and scale. Hassabis said that the AI trajectory is on an “exponential growth curve,” driven by increasing talent, resources, and interest in the field. Hassabis, DeepMind and 60 Minutes are now working on more capable systems that can not only understand language but also the physical environment around them.

This interview was conducted after Google’s Cloud Next 2025 Conference earlier this month. The conference saw the search giant unveil a number of new AI models, features and technologies centered around their Gemini 2.5 multimodal AI family. Google appeared to have taken the lead in providing powerful AI for enterprise applications at affordable prices, surpassing OpenAI.

Details on Google DeepMind’s ‘Project Astra’

The segment focused on Project Astra, DeepMind’s next-generation bot that goes beyond text. Astra is designed for real-time interpretation of the visual world.

It identified paintings, inferred emotions, and created a narrative around a Hopper picture with the line: “Only the flowing of ideas moving forward.”

Astra responded thoughtfully when asked if it had grown bored, revealing an understanding of tone and interpersonal nuance.

Product Manager Bibbo Sh highlighted Astra’s uniqueness: an AI who can “see, listen, and chat about any topic” — a significant step towards embodied AI systems.

Gemini: Towards actionable AI

Also featured in the broadcast was Gemini, DeepMind’s AI system that is being trained to not only interpret the world, but also to act within it – completing tasks such as booking tickets and shopping online. Hassabis stated that Gemini is an important step towards AGI, which is an AI with the ability to navigate complex environments and operate like a human.

The 60 Minutes crew tried out a prototype embedded into glasses, demonstrating real time visual recognition and audio response. Could it also be a hint at a future return of the pioneering but ultimately off-putting early augmented realities glasses known as Google Glass? What debuted in 2012? Being retired in 2015?

Although specific Gemini models like Gemini 2.5 Pro and Flash were not mentioned, Google’s AI ecosystem recently introduced these models for enterprise use which may reflect parallel efforts.

While these integrations support Google’s growing ambitions in AI-based applications, they are outside the scope of the interview.

AGI by 2030? Hassabis, when asked for a timeline of AGI, projected that it could arrive as early as 2030. He said that systems that understand their environment “in very nuanced, deep ways” could be seamlessly integrated into everyday life.

In the interview, Hassabis also discussed the possibility of AI self-awareness. Hassabis stated that current systems are not conscious but future models may show signs of self-understanding. He emphasized, however, the philosophical and biological divide. Even if machines mimic conscious behaviors, they are not the same “squishy matter” as humans.

Hassabis predicted major developments in robots, saying breakthroughs may come in the next few year. The segment showed robots completing tasks based on vague instructions, such as identifying a block of green created by mixing yellow with blue. This suggests that physical systems are becoming more intelligent.

Accomplishments

The segment revisited DeepMind’s landmark achievement with AlphaFold – the AI model that predicted over 200 million protein structures.

Hassabis, along with his colleague John Jumper, were awarded the Nobel Prize in Chemistry in 2024 for this work. Hassabis stressed that this advancement could accelerate drug-development, potentially shrinking timelines by a decade down to only a few weeks. “I think that one day, we may be able to cure all diseases with the help AI,” he said. Hassabis expressed clear concerns despite his optimism. He identified two major risks, the misuse of AI and the increasing autonomy of systems that are beyond human control. He stressed the importance of integrating guardrails and values systems, teaching AI like one would teach a child. He also called for international collaboration, noting that AI will have an impact on every country and culture.

He said that one of his biggest concerns is that the race to dominate AI could become a race down the bottom in terms of safety. He stressed the importance for nation-states and leading players to coordinate on ethical developments and oversight.

This segment ended with a reflection on the future. In a world in which AI tools can transform nearly every human endeavor, they could reshape the way we think about consciousness, knowledge, and the meaning of life. Hassabis said, “We need to have new great philosophers… to understand the implications” of this system.