At the recent Davos Forum, two leading AI executives — Anthropic CEO Dario Amodei and Google DeepMind CEO Demis Hassabis — issued stark warnings about the timeline of artificial general intelligence (AGI) and its profound societal consequences.
Superhuman AGI Could Arrive by 2026-2027
Anthropic CEO Dario Amodei predicted that superhuman-level AGI will emerge between 2026 and 2027. He highlighted the self-accelerating nature of AI development, where programmers primarily edit code generated by AI rather than writing it from scratch. This feedback loop is driving exponential capability growth. “We will soon see machines surpass the smartest humans in every cognitive task,” Amodei stated.
He specifically warned that entry-level white-collar positions — such as administrative assistants, basic data analysts, and customer service representatives — will be the first to face massive displacement. Amodei urged workers in these roles to rapidly upskill or transition to areas requiring human judgment.
DeepMind: 50% Chance of AGI by 2030
In contrast to Amodei’s aggressive timeline, Google DeepMind CEO Demis Hassabis offered a more measured but equally concerning forecast. He said there is a 50% chance of achieving AGI by 2030. Hassabis noted that while scientific creativity remains a challenge for AI — such as autonomously proposing novel theories — the pace of improvement is far faster than expected.
“Scientific creativity is a high bar we haven’t yet cleared, but it’s not an insurmountable obstacle,” Hassabis said. He warned that governments and institutions must begin planning now, or they will be caught unprepared when AGI arrives.
Governments Dangerously Unprepared
Both CEOs agreed that governments are gravely underestimating AI’s impact on labor markets and social structures. Amodei argued that existing social safety nets, education systems, and job training mechanisms are designed for an industrial-era pace and cannot cope with the disruption AGI will bring.
He called for immediate “national-level foresight projects,” including the creation of an AI unemployment compensation fund, redesigning curricula to cultivate uniquely human skills (creativity, empathy, complex decision-making), and establishing international AI safety regulations. Hassabis added that academia and industry must collaborate on new value distribution models to ensure the benefits of technological progress are widely shared.
It is worth noting that both Anthropic and Google DeepMind are among the world’s most advanced AI labs. Anthropic, known for its focus on AI safety, has developed the Claude model series, while DeepMind has achieved breakthroughs in AlphaFold, Go, and beyond. Their warnings are grounded in hands-on front-line research.
As the AGI timeline draws nearer, companies and individual workers must rethink their roles. As Amodei put it: “We’re not trying to stop technological progress — we need to be fully prepared for the world that is coming.”