The general AI arrives: a shock “ten times larger than the industrial revolution, and ten times faster”, alerts the boss of Deepmind

Major figure of theartificial intelligenceDemis Hassabis is nothing of the classic technological leader. Former prodigy of failures, went through research in neuroscience and video games, he heads today Google Deepmindstrategic laboratory of the American giant. Nobel winner for Alphafold, a system capable of predicting the structure of millions of proteins, he says that the next step, general AI, could occur in five to ten years, upsetting economics, science and society.

“Ten times greater, ten times faster”: the prophecy of AI according to the boss of Google Deepmind

The ascent of Hassabis is punctuated by determining choices. In 2010, he co -bonded Deepmind with an ambitious objective: “Resolve intelligence” To meet the major world challenges. Bought by Google four years later, the company became a basic IA search engine. From the victory against the world champion of Go to the publication of the AlphaFold database, his work combines technical feat and concrete applications.

In 2012, a lunch with Elon Musk marked the spirits. The boss of SpaceX says he wants to colonize Mars to ensure the survival of humanity. Hassabis then warns him: “And if what was wrong, it was AI? Being on Mars wouldn’t help you.”

The dazzling boom in generative models, as Chatgpthas since changed the situation. Google now integrates AI to all its serviceswhile competition – OpenAi of course, but also Meta, Microsoft, Apple – multiplies investments. This global race accelerates the deadline for this general artifice intelligence, or AG, capable of reproducing all human cognitive capacities.

We should enter a world of radical abundance, with unprecedented medical, energetic and scientific breakthroughs.

This perspective enthusiasm as much as it worries. On the scientific level, Hassabis imagines major advances: nuclear fusion, new materials, revolutionary treatments … But it recognizes the risks: concentration of power, Massive job replacementhigh energy consumption of data centers. For him, the key question will not only be technological but political: how to redistribute the profits fairly?

He also sees a cultural issue: in a world where the production of material goods would be largely automated, it would be necessary to restore its place to non -utilitarian human activities – arts, sport, philosophy. This tilting would require a collective effort to rethink the meaning and purpose of the work.

Hassabis says “Prudent optimistic”. Industrial history has shown that such transitions are sources of progress but also strong inequalities. The challenge, he concludes, will be to guide this transformation to make it a positive step in human history, rather than an irreversible break.