Artificial intelligence is rapidly transforming how we live, work, learn, and interact with one another, and government is no exception. Too often, however, the focus is on the capabilities of the technology rather than the challenges it could help overcome. The question facing public sector leaders is not whether to embrace AI, but how to harness its potential to solve real problems and improve the lives of the people they serve.
The technology holds substantial promise to speed up services, clear backlogs, and return staff time to work that needs human judgment.
Virtually every state has an AI pilot under way, yet whatâs yet unclear is if theyâve been fully developed to measure public value. Early in a technologyâs life it is easy to change course and hard to know the consequences, and by the time the consequences are clear, the systems, contracts, and habits are locked in. Government cannot wait for conclusive evidence before it acts, so it must decide in advance what any use of AI will be required to demonstrate. The obstacle the efficiency assumption, the belief that AI is efficient by nature and therefore needs no proof. Efficiency is a claim to be tested for each use, and even proven efficiency is not enough if people cannot see, challenge, or reverse what the system does.

