Google DeepMind CEO Sir Demis Hassabis called for urgent research into AI threats at the AI Impact Summit in Delhi, warning that the industry needs "robust guardrails" against bad actors and loss of control over increasingly powerful systems. Hassabis acknowledged that keeping pace with AI development is "the hard thing" for regulators, and said his firm was "only one player in the ecosystem." The summit, attended by delegates from over 100 countries, saw broad calls for global AI governance — though the US delegation explicitly rejected that approach, with White House adviser Michael Kratsios dismissing centralised control of the technology.

Urgent research needed to tackle AI threats, says Google AI boss - BBC News