Google DeepMind CEO Demis Hassabis has reiterated his prediction that artificial general intelligence (AGI) — a theoretical AI capable of matching human-level cognition — will be achieved within the next few years.
Speaking at Stanford Graduate School of Business, Hassabis said he believes AGI is roughly three to four years away, estimating a timeline of “maybe 2030, plus or minus a year.” He described the arrival of AGI as “an enormous transformative technology” that would mark the beginning of “a new human era.”
Looking further ahead, Hassabis suggested that within a decade, society will recognize it was “standing in the foothills of the singularity” — the hypothetical point where AI surpasses human intelligence and begins improving itself autonomously, beyond human control. The remarks come amid intensifying industry debate about the pace and implications of advanced AI development.
There is no consensus on AGI’s arrival even among leading AI executives. Sam Altman of OpenAI claimed last year in a blog post that “humanity is close to building digital superintelligence,” a hypothetical stage beyond AGI where AI exceeds the smartest human minds. He noted that the entire industry, not just his company, is engaged in that effort.
Hassabis’s comments add to a growing chorus of tech leaders offering concrete timelines for AGI, even as critics warn such predictions risk overhype or downplaying safety risks. The singularity remains a divisive concept, with some experts arguing it misrepresents how AI will evolve.