Google DeepMind CEO Demis Hassabis says Artificial General Intelligence, or AGI, AI with human-like cognitive abilities, could be just a few years away, calling it the beginning of a new era for humanity.

He believes AGI could transform fields like medicine, clean energy and scientific research at an unprecedented pace.

But he also warns that AGI could pose serious risks, involving cybersecurity, biological and nuclear threats, if it is developed without adequate safeguards.

Hassabis proposes creating a US-led Frontier AI Standards Body to test the most advanced AI models, establish safety benchmarks and encourage responsible development.

He stresses that these evaluations should become mandatory before frontier AI models are deployed for wider use.

He proposes that the framework should eventually form the basis for international AI standards, ensuring AGI benefits society while its biggest risks are effectively managed.

Published on July 16, 2026