Why does Karen Hao question the goal of building AGI?

Karen Hao says the push for AGI rests on an unproven hypothesis held by researchers such as Ilya Sutskever and his mentor Geoffrey Hinton: that the brain is a statistical engine, so ever larger statistical models will match and then exceed human intelligence. Hao argues that this bet drives more data collection, more data centres and more labour exploitation, and that technology should improve human flourishing rather than replace people.

• Hao recalls a Sutskever keynote at the NeurIPS conference showing a roughly linear chart of brain size against species intelligence.
• As an alternative, Hao points to narrower AI for drug discovery and better healthcare outcomes.

▸In their words

“ultimately our brains are giant statistical models.” Karen Hao · at 4:01 — open in the transcript →

Source: our own transcription of the episode. The full conversation is on The Diary Of A CEO with Steven Bartlett (2026-10-23). Read the whole episode summary →

▸More from this episode