Google AI CEO Demis Hassabis shares three reasons why Sam Altman's 'ultimate goal' for OpenAI cannot be realised

Google DeepMind CEO Demis Hassabis has said that artificial general intelligence (AGI) still remains far away. He has outlined three limitations that, in his view, will prevent OpenAI chief Sam Altman’s ‘ultimate goal’ of achieving AGI from being realised in the near future. Speaking at the India AI Impact summit in New Delhi, Hassabis said current systems do not yet match human intelligence. “I don’t think we are there yet,” he said.


His comments come as OpenAI continues to frame AGI as its long-term objective. Altman has said that more advanced forms of intelligence could follow, arguing that “Superintelligent tools could massively accelerate scientific discovery and innovation well beyond what we are capable of doing on our own, and in turn massively increase abundance and prosperity.” Last year, OpenAI also restructured itself as a Public Benefit Corporation as part of efforts to align its development work with broader societal goals.

Why Google AI CEO Demis Hassabis thinks AGI is not ready yet

Hassabis explained that today’s AGI-style systems lack key capabilities, starting with what he described as “continual learning”. He said these models remain largely fixed after training and do not adapt in real time to new experiences or contexts.

“What you’d like is for those systems to continually learn online from experience, to learn from the context they’re in, maybe personalise to the situation and the tasks that you have for them,” he explained.

Secondly, Hassabis said current systems have limitations in long-term reasoning. “They can plan over the short term, but over the longer term, the way that we can plan over years, they don't really have that capability at the moment,” he said. Hassabis also said these systems lack consistency, performing well in some areas while falling short in others.

“So, for example, today's systems can get gold medals in the international Math Olympiad, really hard problems, but sometimes can still make mistakes on elementary maths if you pose the question in a certain way,” he said.

“A true general intelligence system shouldn't have that kind of jaggedness,” Hassabis added. For comparison, human experts would not typically make errors on basic problems in their area of expertise, he noted.

Last year, in a “60 Minutes” interview, Hassabis said that he expects true AGI to emerge within five to 10 years. In 2010, he co-founded the AI research lab DeepMind, which Google acquired in 2014 and is responsible for developing the tech giant’s Gemini models. In 2024, Hassabis received a joint Nobel Prize in chemistry for his work on protein structure prediction.

AGI remains a debated concept in Silicon Valley. Databricks CEO Ali Ghodsi said at a conference in September that current AI chatbots already meet the definition of AGI, arguing that industry leaders keep “moving the goalposts” by shifting the focus toward superintelligence, or AI that can outperform humans.

The AI Summit in India, taking place this week, has drawn speakers from across the technology and AI sectors. The agenda includes figures such as OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, Google CEO Sundar Pichai, and Meta’s chief AI officer, Alexandr Wang.