DeepMind CEO Hassabis Sets AGI Few Years Out, Urges Global Safety Standards

Demis Hassabis, CEO of Google DeepMind, projected that artificial general intelligence could arrive within the next few years and urged governments and industry to establish global safety standards before the technology emerges.

The announcement aligns with his recent public statements narrowing AGI timelines from an earlier 5-10 year range to roughly 2029-2030.

Hassabis defined AGI as AI matching or exceeding human intelligence across scientific creativity and continuous learning. Multiple interviews place the updated horizon at three to five years or around 2029-2030. He noted that current systems show jagged intelligences and that remaining challenges include continual learning, long-term reasoning, memory, and world models grounded in physics and causality. His remarks coincide with Google DeepMind releases of Gemini 3.0 and 3.5 models plus new agentic products at the Google I/O developer conference.

The tightened timeline builds on earlier forecasts. CNBC reported Hassabis stating human-level AI would arrive in 5 to 10 years. The Times of India quoted him saying AGI remains 5-10 years away and requires one or two additional major breakthroughs, while maintaining a 50 percent chance by 2030. Axios coverage of his post-Google I/O remarks described 2029 as a real possibility and labeled the language intentionally provocative toward governments and economists. Business Insider quoted his Stanford Graduate School of Business fireside chat: Maybe 2030, plus or minus a year, which is astounding to think, really. I think that will be such an enormous transformative technology; it’s gonna effectively be a new human era.

The shift occurs as DeepMind deploys multi-agent systems such as Co-Scientist across 17 U.S. Department of Energy national labs. Google co-founder Sergey Brin separately predicted AGI before 2030, reinforcing the accelerated outlook. Hassabis framed AI agents as an intermediate stage that will manage other systems with humans increasingly out of the loop, projecting 500-1,000 times productivity gains for engineers compared with the early 2000s.

These forecasts increase pressure on regulators worldwide to coordinate safety frameworks before AGI-level systems appear. African governments including Nigeria’s NITDA, NCC, and Ministry of Communications, Innovation and Digital Economy face urgency to engage with emerging global standards rather than await mature Western rules. If standards are set primarily by US, EU, and UK bodies, local startups may encounter compliance burdens designed for richer ecosystems when exporting AI products or integrating with global platforms.

Frontier access and dependence also shift. Google Gemini and DeepMind technologies already integrate into cloud platforms used by African startups. Rapid AGI arrival could deepen reliance on a small set of foreign labs. Policymakers must negotiate for access, data protections, and localisation while reskilling youth-heavy, services-oriented economies for massive labour changes. African researchers and civil society groups working on AI ethics and bias gain a window to ensure global standards incorporate Global South perspectives on data scarcity and differing social risks.

Maybe 2030, plus or minus a year, which is astounding to think, really. I think that will be such an enormous transformative technology; it’s gonna effectively be a new human era.

Nairametrics coverage shows Nigerian financial and tech media tracking these frontier-lab claims, shaping local investor and founder expectations toward AI-first and agentic products.

Observers should watch responses from African regulators and whether Global South voices influence upcoming international AI governance discussions. Google DeepMind’s next model releases and agent deployments will provide further signals on whether the 2029-2030 window holds.