**Pau Barrena | AFP | Getty Images** LONDON — The full realization of artificial intelligence (AI) that matches human capabilities is still on the horizon, but Google DeepMind's CEO, Demis Hassabis, believes that artificial general intelligence (AGI)—intelligence that equals or surpasses human cognition—will start to emerge in the next five to ten years. During a briefing in London, Hassabis emphasized that while current AI systems excel in certain areas, they remain limited in others, and substantial research is needed before achieving true AGI. He described AGI as a system capable of exhibiting the complex skills that humans possess. Hassabis' timeframe for AGI contrasts with predictions from other industry figures. Baidu’s CEO, Robin Li, noted last year that AGI is "more than 10 years away, " while Anthropic's Dario Amodei optimistically suggested it could emerge in "two to three years. " Cisco's Chief Product Officer, Jeetu Patel, went even further, forecasting signs of AGI could appear as early as this year. After AGI, artificial superintelligence (ASI) is anticipated, though the timing of its arrival remains uncertain.

Musk and OpenAI's Sam Altman have also made predictions about the timeframe for AGI, with Altman suggesting it could emerge in the "reasonably close-ish future. " Hassabis pointed out that a significant challenge in achieving AGI is enabling AI systems to understand real-world context. While AI has excelled in specific domains like games (for example, Go), translating this capability to the real world is more complex. He noted advancements in "world models" and the need for effective integration with planning algorithms. Hassabis and Google Cloud CEO Thomas Kurian highlighted the promise of "multi-agent" AI systems, which are evolving behind the scenes. They are working on enabling AI agents to play complex games like "Starcraft, " promoting cooperation and communication among agents to enhance their functionality and interaction.