Google DeepMind CEO warns AGI could arrive in a few years, calls for international AI safety review

Google DeepMind CEO Demis Hassabis has warned that highly capable artificial general intelligence (AGI) could be just a few years away, urging society to prepare faster. He said on social media there is still a 'precious window' to act before AGI arrives.

Hassabis suggested forming an international coalition led by the United States to review advanced AI models before public release. He proposed a model similar to Wall Street's self-regulatory body, funded by the industry, with flexible and rigorous testing methods.

Hassabis previously predicted AGI could arrive by 2030 and has repeatedly said society is not ready. He believes experts from philosophy, economics and other fields should help assess AI's deep social impacts. He stressed that choices made now will shape the next phase of civilization.