AGI has become mainstream as a topic of discussion only recently, but Google Deepmind’s co-founder has been predicting AGI timelines since 2009 — and he hasn’t changed his prediction.
Shane Legg, who co-founded what is now Google DeepMind alongside Demis Hassabis and Mustafa Suleyman, has maintained a remarkably consistent stance on artificial general intelligence timelines for over 15 years. While the AI industry has cycled through periods of hype and skepticism, Legg’s forecast has remained steady: a 50% probability of AGI arriving by 2028. What makes his recent comments particularly striking isn’t just the timeline itself, but his emphasis on the profound societal transformation that awaits us.
“I’ve publicly held the same prediction since 2009: there’s a 50% chance we’ll see AGI by 2028,” Legg says. “This is actually something which is going to structurally change the economy and society and all kinds of things. And we need to think about how do we structure this new world.”
His vision extends beyond the technological achievement itself. Legg sees AGI as a potential catalyst for unprecedented prosperity, though not without careful consideration of its implementation. “This could be a real golden age because we now have machines that can dramatically increase production of many types of things,” he explained.
However, Legg was quick to emphasize that technological capability alone doesn’t guarantee positive outcomes. “So there’s an opportunity here, but that is only good if we can somehow translate this incredible capability of machines into a vision of society where there is some flourishing of people as individuals and as groups of people in society that benefit from all this capability.”
Legg’s consistent prediction takes on new significance in light of recent developments across the AI industry. OpenAI CPO Kevil Weil has said that AGI could arrive before 2027, while Anthropic CEO Dario Amodei has pointed to 2026 or 2027 as plausible timelines. Softbank’s Masa Son believes it can arrive much sooner. Meanwhile, DeepMind itself continues to push boundaries with systems like Gemini and AlphaFold, demonstrating increasingly general capabilities across domains from language to scientific research. The convergence of these predictions from leading AI researchers, combined with rapid advances in model capabilities, suggests that what once seemed like distant speculation is now being treated as an imminent reality requiring serious policy consideration. Legg’s focus on societal structure and distribution of benefits reflects a growing recognition among AI leaders that the technical challenge of creating AGI may ultimately prove easier than the social challenge of ensuring it benefits humanity broadly.