Lord Gus O’Donnell, who led Britain’s civil service from 2005‑2011, told the Downing Battcock Institute audience that the rapid spread of artificial intelligence will slash public‑sector costs but also push large numbers of employees out of work. He urged that companies reaping AI‑driven gains should shoulder the burden of retraining displaced staff through a dedicated tax levy.

O’Donnell argued the extra revenue would flow from higher corporation tax on AI‑enhanced profits and increased income tax as workers move into higher‑paid roles. He singled out digital ID as the linchpin for unlocking AI’s potential, insisting that a secure, government‑issued identifier would enable safe data sharing and justify the productivity surge investors seek, and would fund sector‑wide reskilling initiatives.

The former Treasury permanent secretary warned that without a clear financing framework, the social cost of AI could erode public support for digital reforms. He called on ministers, including Chancellor Rachel Reeves, to undergo AI training themselves and to experiment with tools like Claude or ChatGPT for drafting policy speeches. Implementing his proposal would channel new tax receipts into a systematic upskilling pipeline.