The UK's real AI policy is procurement

Public procurement has become the de facto AI policy in the United Kingdom. The Open Rights Group's August report on UK public-sector AI looks at first glance like a complaint about Big Tech; read as evidence, it is a structural diagnosis. Contracts flow to a narrow set of large US-headquartered technology firms that supply AI for benefits, policing, immigration, and health triage, and those same firms then lobby Whitehall for further roll-out. The pattern is procurement dependency feeding a lobbying feedback loop, with ministers holding the procurement lever and the public holding the consequences.

ORG frames adoption in its own terms: opaque, unaccountable, and built for Big Tech rather than the public. The group also flags foreign-vendor exposure, warning that reliance on "foreign capital and systems designed and operated by private companies already boasting market dominance" leaves the public sector dependent on actors in jurisdictions the UK does not control. Three MPs, from Labour, the Liberal Democrats, and the Green Party, wrote forewords endorsing regulation.

Vendors who set the menu set the policy. Where ORG is right, the next contract is the policy event. Where the report is thin, it is on decision-stage harm, since procurement opacity does not prove case harm. Independent scrutiny, not the report, would settle that.

The address sits in one place. The Minister for AI, Kanishka Narayan, can keep procurement as a backdoor AI policy, or open it before the next contract lands.