Embedding AI in Public Services

The government wants to accelerate the digital transformation of public services and has recruited several academics in the field to advise.

Facebook owners Meta have made a $1m investment in the program.

The aim is to use AI “to make public services faster, safer and more efficient”, and to “re-wire” systems across healthcare, policing, transport and more.

In a separate project AI developer Anthropic are developing a system to assist jobseekers with tailored career advice and practical assistance. Collaboration is essential to deploy AI “safely for the public benefit.”

Extracts from Government Transformation Magazine

These initiatives are quite separate from the contacts already awarded to US giant Palantir the UK government’s most significant and controversial partners in its digital transformation journey. Palantir are working on systems in the NHS, MoD, Policing and Border Control. 

There are significant concerns here in that these companies aren’t “partners” who can easily have their contracts terminated in favour of another supplier. Their systems and people become embedded as part of the fabric of public service provision. They effectively become the private owners of public services. Once we hand over control to companies with a monopoly on service provision and whose sole motive is profit, the government has lost all control.

Many contracts have been agreed via “direct award” (without a competitive tender process) under “national security” or “emergency” exemptions, leading to accusations of cronyism and a lack of democratic oversight.

We need to proceed with extreme caution, but if PFI is our guide, there is little room for optimism. The long-term financial risk, “naive” procurement, and a loss of public control, are remarkably similar.