Google DeepMind is dedicated to achieving artificial general intelligence (AGI), aiming for AI with human-level cognitive abilities but with enhanced speed and knowledge, according to CEO Demis Hassabis in a 60 Minutes interview. Hassabis envisions AI becoming deeply integrated into daily life, solving complex problems and even displaying imagination.

Hassabis, a pioneer in the field and co-founder of DeepMind (acquired by Google in 2014), has always been driven by a desire to understand the world. His academic background includes studies at Cambridge, MIT, and Harvard, culminating in a doctorate in neuroscience. Hassabis views AI as a powerful tool for advancing human knowledge, exemplified by DeepMind's AI model that predicted the structure of 200 million proteins in a single year, a task that would traditionally take years to accomplish manually.

AI's rapid progress is astounding, even to those within the field. DeepMind's Project Astra, an AI companion, can see, hear, and engage in conversation about a wide range of topics. During a demonstration with 60 Minutes, Astra identified artworks and even created a narrative about Edward Hopper's "Automat," showcasing the AI's ability to interpret and understand the world around it.

Hassabis notes that AI systems often surprise developers with their capabilities, as they learn independently from vast amounts of internet data. While developers establish the foundational architectures, the specific skills and properties emerge from the training process, akin to how humans learn. DeepMind is currently training its AI model "Gemini" to perform practical tasks such as booking tickets and online ordering. Robotics is also a key area of advancement, with Hassabis predicting breakthroughs in humanoid and other robots capable of performing useful tasks within the next few years.

While AI is not yet self-aware or conscious, Hassabis acknowledges the possibility of these developments in the future but emphasizes focusing on building intelligent tools first. He also points out that AI currently lacks imagination, as these systems primarily operate based on existing human knowledge. They can't formulate truly novel questions or hypotheses.

Despite these limitations, Hassabis believes AI offers enormous potential benefits. He anticipates AI could revolutionize drug development, potentially reducing the time and cost of creating new drugs from years to months or even weeks. He also envisions AI leading to “radical abundance,” effectively eliminating scarcity. However, Hassabis stresses the importance of implementing safety measures and guardrails to prevent misuse and ensure control as AI systems become more powerful.

Hassabis advocates for international cooperation among leading players and countries to establish safety limits and ethical guidelines for AI development. He suggests that AI systems can be taught morality by demonstrating and teaching values, similar to how children are raised. Ultimately, he is confident that AI, when developed responsibly, can be a force for good, offering solutions to some of humanity's greatest challenges.