AI in government – how to categorise the opportunities?

The UK government's Incubator for AI (i.Ai) - a government team of technical experts who are tasked to help departments harness AI to improve lives and improve delivery of public services - has grouped together different categories of AI projects in UK government. This is known as the i.Ai taxonomy (here).

i.Ai's categories are:

1. Public facing services - the interface between government and its citizens. This does not necessarily mean the AI is used to produce output that goes directly to the citizen. The example in the i.Ai's taxonomy is of a tool called ‘Caddy’ - a customer service AI copilot that helps advisors and supervisors quickly locate and share information from reliable sources.
2. Fraud and error - through identifying anomalies and analysing patterns in data which may indicate fraud or error.
3. Matching and triage - by analysing for patterns, helping to match resources to public sector needs, such as hospital equipment, or matching grid capacity and generation. 
4. Casework management - through processing, synthesising and analysing paperwork faster than a human could, reducing the time required and need to outsource, but ultimately augmenting (rather than replacing) current work.
5. Data infrastructure - i.Ai give the example of their Project rAPId, an open-source solution for departments to discover, manage and share data internally and across government.

i.Ai's approach is not to classify a project by policy area - for example, tied to a specific department - but instead by 'user and technical challenge'.  i.AI says that this reflects that often there are similar solutions to similar problems in these categories. The aim is to focus AI development on solutions to technical problems, not policy problems. 

If you would like to discuss how current or future regulations impact what you do with AI, please contact Tom Whittaker, Brian Wong, Lucy Pegler, David Varney, or Martin Cook. For the latest on AI law and regulation, see our blog and newsletter.

The opportunities for AI in government are vast. The Alan Turing Institute assesses that 84% of the 143 million complex, repetitive transactions that take place across government services every year are ‘highly automatable’ (The Alan Turing Institute, AI for bureaucratic productivity). The Tony Blair Institute calculates that through embracing AI, the UK stands to gain around £40 billion in public sector productivity improvements (Tony Blair Institute, Governing in the Age of AI). The size of the prize is both exciting and intimidating. The pressure to pick the right projects from a never-ending list can seem high.