AI matching human intelligence ‘five to ten years away’

• DeepMind CEO Demis Hassabis predicts the emergence of artificial general intelligence (AGI) within the next five to ten years, followed by artificial superintelligence.
• There are varying timelines for AGI, with some leaders believing it is more than 10 years away and others suggesting it is imminent within two years.
• World leaders have convened to establish guidelines for AGI development, highlighting potential risks and the possibility of dominance by a few companies and countries.

---

Full Story

At a DeepMind briefing on Monday, March 17, CEO Demis Hassabis said today’s artificial intelligence systems are passive and lack certain capabilities. He added that in the next five to ten years, artificial general intelligence (AGI) — systems that can perform all human tasks — could emerge.

Hassabis noted that artificial superintelligence is expected to follow AGI, but the timeframe for its arrival remains uncertain. This aligns with predictions from Baidu CEO Robin Li.

“Today a lot of people talk about AGI. They’re saying it’s probably two years away, it’s probably five years away,” Li said at a conference in May 2024. “I think it’s more than 10 years away.”

However, other tech leaders from companies like Cisco and Anthropic believe AGI could emerge within the next two years.

The varying timelines have prompted world leaders to convene for a major summit in Paris in February 2025. Executives and experts aimed to establish guidelines for the rapid development of the technology, with scientists issuing warnings about its potential risks.

“The few companies and countries developing AGI will have the ability to dominate other businesses with AI-driven operations,” Yoshua Bengio, the Scientific Director at Mila Quebec AI Institute, said. “These superhuman A.I.’s may not be shared globally.”

Hassabis identified the primary challenge in achieving AGI as advancing current AI systems to understand real-world context.