Google DeepMind’s Pushmeet Kohli, at HT Leadership Summit 2025, said the lab’s mission is to “push the boundaries of AI” and build the most powerful, useful models possible while keeping safety and societal benefit at the core of its work. He spoke about advancing frontier capabilities in reasoning and science, but stressed rigorous evaluation, alignment research and strong guardrails so that next‑generation AI systems remain trustworthy and broadly beneficial rather than disruptive.
E-Paper
