Google DeepMind founder and Nobel laureate Demis Hassabis has shortened his estimated timeline for artificial general intelligence to just two to three years. Defining AGI as a system that matches the full cognitive range of the human brain, Hassabis noted that his projection continues to move closer to the present. Earlier this year, he had estimated the arrival of AGI within three to five years, reflecting a rapidly accelerating trajectory in artificial intelligence development.
To measure when we have reached true AGI, Hassabis relies on what he calls the Einstein test, asking whether a system trained exclusively on pre-1901 knowledge could autonomously reproduce Albert Einstein's groundbreaking 1905 physics papers. Having previously compared the potential impact of AGI to the industrial revolution happening at ten times the speed, Hassabis now focuses on Isomorphic Labs, Google's drug discovery AI spin-off. He has also advocated for an industry-funded oversight board operating under government control to safely vet advanced AI models before public release.
This updated timeline from a leading AI pioneer highlights the rapidly accelerating pace of advanced artificial intelligence research.