Experts Warn AGI May Arrive by 2029, Urge Societal Preparation

Advanced large language models now power AI chatbots that can write, reason, and interact in ways that often seem human. As these systems advance, AI experts are focusing on the next stage: Artificial General Intelligence (AGI). At the recent Google I/O developer conference, Demis Hassabis, CEO of Google DeepMind, stated that humanity stands in the “foothills of the singularity.” He warned that society may have only a few years to prepare for AGI.

Key Highlights

• Demis Hassabis warns AGI could arrive as early as 2029.
• Experts say society has only a few years to prepare for AGI.
• AI industry lacks consensus on the definition of AGI.
• Current AI systems show signs of soft self-improvement.
• AGI

AGI Timeline and Expert Warnings

Hassabis said he broadly expects AGI around 2030 but now sees 2029 as a real possibility. Speaking to Axios after Google I/O, he noted, “We can see agents really happening now and imagine what they will be in another year, and how useful they'll be.” Hassabis stressed the need for faster preparation by governments and society. He said he used dramatic language to increase urgency among policymakers, economists, and the public. According to Hassabis, most discussions about AI’s impact remain within technology circles, even though the technology could affect nearly every aspect of life.

Milestones and Self-Improvement in AI

One major milestone for researchers is the potential for AGI-level systems to achieve recursive self-improvement. This means AI could begin improving its own capabilities at an accelerating pace. Hassabis told Axios that this is a key milestone being monitored by leading labs. “All the leading labs are quite focused on that,” he said. However, he clarified that current systems are not yet fully improving themselves independently. “We're not yet at the point where the systems are getting better on their own, but the pace of development is clearly accelerating,” Hassabis explained.

He described today’s coding agents as a form of “soft self-improvement.” These agents already make engineers significantly more productive, but they do not yet represent full autonomous self-improvement.

Debate Over AGI Definition

Even as AI companies race toward AGI, there is no industry-wide agreement on what qualifies as AGI. Sam Altman, CEO of OpenAI, addressed this in a February 2025 blog post. He wrote that AGI is “a weakly defined term,” but OpenAI considers it a system capable of solving increasingly complex problems at a human level across many fields. Altman stated, “In some sense, AGI is just another tool in this ever-taller scaffolding of human progress we are building together. In another sense, it is the beginning of something for which it’s hard not to say ‘this time it’s different.’”