Google DeepMind CEO Demis Hassabis warned that AI's rapid advancement could outpace safety measures, urging proactive governance as the technology enters a transformative phase. Speaking at the AI Impact Summit in New Delhi, Hassabis highlighted dual-use risks: bad actors exploiting AI for cyberattacks or bioweapon development. He emphasized that current systems already excel at cyber threats, demanding stronger defenses to prevent catastrophic misuse. AGI (Artificial General Intelligence) could accelerate these dangers, with autonomous agents potentially acting unpredictably within 12-18 months, he noted.

The summit, hosted in India for the first time, aims to center developing nations in global AI policy debates. Officials stress that equitable governance frameworks are critical to prevent power imbalances between tech giants and emerging economies. Hassabis’ remarks align with broader concerns about AI’s dual-use nature—its capacity to revolutionize science while enabling unprecedented harm if unregulated.

Cybersecurity and biosecurity emerged as top priorities, with Hassabis warning that AI’s growing capabilities in these fields require immediate safeguards. He called for collaboration between governments and companies to establish “red lines” before AGI deployment. The summit’s focus on inclusive governance reflects growing recognition that AI’s risks transcend borders, demanding multilateral solutions.

This AI safety debate underscores a pivotal moment: balancing innovation with accountability. As India positions itself as a global AI voice, Hassabis’ warnings serve as a reminder that unchecked progress could undermine the very trust needed to sustain the technology’s benefits.