Google DeepMind Chief Urges Urgent AI Risk Research
The head of Google DeepMind, Sir Demis Hassabis, has called for urgent research into the risks posed by artificial intelligence, warning that safeguards must keep pace with rapidly advancing technology.
Speaking in an interview on the sidelines of the AI Impact Summit in New Delhi, Hassabis said the industry supported “smart regulation” that focuses on the most serious dangers associated with AI, rather than broad restrictions that could stifle innovation.
He identified two major threats: the misuse of AI by malicious actors and the possibility of humans losing control over increasingly autonomous systems. According to him, building strong guardrails around advanced AI models is becoming more urgent as the technology grows more powerful.
Hassabis acknowledged that technology firms alone cannot slow development to give regulators time to respond, describing DeepMind as “only one player in the ecosystem”. He added that keeping up with the speed of innovation remained a major challenge for policymakers worldwide.
Calls for stronger oversight echoed across the summit, which brought together delegates from more than 100 countries. Sam Altman, head of OpenAI, also urged urgent regulation, while India’s prime minister, Narendra Modi, stressed the need for international cooperation to ensure AI benefits society.
The United States, however, pushed back against proposals for global governance. White House technology adviser Michael Kratsios said Washington rejected centralised control of AI, arguing that excessive bureaucracy could undermine its potential.
Hassabis said western countries currently hold a slight lead over China in AI development but warned the gap could close within months. He added that those building advanced systems must balance ambition with responsibility.
Looking ahead, the DeepMind chief said AI could become a “superpower” for human creativity over the next decade, while emphasising that science and technology education would remain important in an AI-driven world.
The summit is due to conclude with discussions on shared principles for managing artificial intelligence, though deep divisions remain over how far regulation should go.
