DeepMind’s CEO, Demis Hassabis, has identified three key areas where current artificial general intelligence (AGI) systems still fall short of human-level intelligence. Speaking at an AI summit in New Delhi, Hassabis acknowledged that while progress is rapid, true AGI has not yet been achieved. When asked whether today’s AGI systems can match human intelligence, Hassabis responded clearly: “I don’t think we are there yet.” He outlined three critical limitations holding back the development of systems that can truly reason and adapt like humans. The first challenge is continual learning. Unlike humans, who constantly learn and adapt from real-world experiences, current AI systems are largely static after training. “What you'd like is for those systems to continually learn online from experience, to learn from the context they're in, maybe personalize to the situation and the tasks that you have for them,” Hassabis explained. The second area of weakness is long-term planning. While AI can make short-term decisions and solve immediate problems, it struggles with strategic thinking over extended periods. “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. The third issue is inconsistency in performance. AGI systems can excel in highly complex tasks—such as solving advanced math problems at the level of an International Math Olympiad—but still fail on basic arithmetic when presented in a slightly different format. “A true general intelligence system shouldn't have that kind of jaggedness,” Hassabis noted. He contrasted this with human experts, who would not make elementary mistakes in their area of expertise. Hassabis has previously stated in a “60 Minutes” interview that he expects true AGI to emerge within five to ten years. As co-founder of DeepMind, established in 2010, he has led groundbreaking work in AI, including the development of AlphaGo and the protein structure prediction tool AlphaFold, which earned him a joint Nobel Prize in Chemistry in 2024. The debate over what constitutes AGI remains active in Silicon Valley. At a September conference, Databricks CEO Ali Ghodsi argued that current AI chatbots already meet the definition of AGI, suggesting that industry leaders are simply raising the bar to justify further investment in superintelligence—systems that could surpass human cognitive abilities. The AI Summit in India, which ran from Monday to Friday, brought together top figures in the field, including OpenAI’s Sam Altman, Anthropic’s Dario Amodei, Google’s Sundar Pichai, and Meta’s chief AI officer Alexandr Wang. The event highlighted the growing global focus on AI’s future and the challenges of building systems that truly mirror human intelligence.