Google DeepMind: Artificial General Intelligence Could Become a Reality in 5-8 Years

Artificial general intelligence (AGI) - systems capable of reasoning similar to humans - could be developed in the next five to eight years, according to statements made in New Delhi by Google DeepMind CEO Demis Hassabis, at an international summit dedicated to the impact of AI technology. "We are 5-8 years away from reaching an artificial intelligence that can reason similarly to a human being," he said, according to the EFE agency, underlining the accelerated pace of progress in the field.

• "Einstein Test", new standard proposed for evaluating AI

To measure the real innovation capacity of intelligent systems, Hassabis proposed a new criterion, called the "Einstein Test". The concept involves training an AI system with all of human knowledge, but with data artificially frozen in 1911. Subsequently, researchers would analyze whether the algorithm could discover on its own the theory of general relativity, developed in 1915 by Albert Einstein. According to Hassabis, this type of test would differentiate between systems that simply reproduce existing information and those capable of generating original knowledge.

"Current tools are encyclopedic experts, but they do not yet have a real creative dimension. They solve what already exists and do not formulate revolutionary scientific hypotheses,” he explained.

• Current limits and the road to AGI

The director of DeepMind showed that achieving general intelligence will require the combination of several technologies: the strategic planning capacity demonstrated by systems such as AlphaGo; the processing power of modern fundamental models; the integration of Gemini-type models, which provide a complex representation of how the world works. This evolution would transform AI from a rapid information access tool into a "discovery engine” capable of solving scientific problems that currently exceed human capabilities.

• Global summit on the future of AI

The statements were made at the India AI Impact Summit 2026, considered the largest global event dedicated to artificial intelligence to date. The event brings together more than 20 heads of state and about 500 industry leaders, with the aim of defining the rules that will govern the digital economy in the next decade and the balance between innovation and security.