On 1 October 2025 the Francis Crick Institute published Can we harness AI for good?, an episode of A Question of Science recorded for the Crick with BBC Studios. Professor Brian Cox chaired; the panel was Adrian Weller (Cambridge and the Alan Turing Institute), Jeanette Winterson, Steph Wright (Scottish AI Alliance) and Neil Lawrence.

Asked how AI already shapes daily life, Neil argued that the functional core — machines that recognise and reconstruct patterns in data — has been embedded for years in advertising, search and navigation, and that the present hype comes from large-scale systems that emulate conversation. The hard work, he said, is to move past a conversation driven by claiming space and making money toward one about who we are and who we want to be. He returned to the bandwidth gap developed in The Atomic Human: machines absorb information orders of magnitude faster than people can communicate, while evolutionary and climatic systems run far slower — so machines that decide alone will struggle to understand us, and need guidance from people.

On artificial general intelligence, he treated public fears as valid but misplaced if they ignore the present: digital systems already undermine livelihoods when access to the information infrastructure is restricted to a few companies — the Horizon scandal being a case in point. The episode page includes the full transcript.