DeepMind CEO Says AGI Is 5–8 Years Away, Calls It A Coming ‘Golden Era’

Artificial intelligence (AI) has come a long way in a short time. From solving complex scientific problems to writing code and cracking tough exams, AI systems are pushing boundaries. However, when it comes to Artificial General Intelligence (AGI), machines that can think and learn like humans across tasks, the finish line is still ahead. Well, that’s essentially what the conversation at the India AI Impact Summit came down to.

 

TL;DR

 

• Demis Hassabis says AGI is on the horizon but not here yet.
• AI still struggles with consistency, long-term planning, and continued learning.
• He estimates a 5–8 year window for reaching a key AGI threshold.

Speaking at the India AI Impact Summit in New Delhi, Google DeepMind CEO Demis Hassabis offered a clear-eyed view of where things stand. We aren't there yet in terms of Artificial General Intelligence, or AGI, but it's certainly on the horizon, he said during a conversation with IIT Madras Professor B. Ravindran.

“We are going to enter a golden era for scientific discovery, almost a new era,” Hassabis said, reflecting on how AI is reshaping research. The Nobel Prize-winning scientist, recognized in 2024 for his work on protein structure prediction using AI, sounded optimistic but grounded.
 
As for AGI, “we are not there yet. “There are still things that systems can't do — things like continued learning still needs to happen,” he explained.
 
He highlighted a key gap of reliability saying “Systems today can get international Math Olympiad medals, but can still make mistakes in elementary math. (AI) is still jagged that way.”

Still, the progress has been rapid. “It's been amazing to see the progress we have made in just over a decade. The threshold moment where AGI is on the horizon is 5-8 years,” he added.

For now, AGI remains a goal in sight, but not yet within reach.