Public Registers of AI Systems

If you wanted to find out which AI systems your local council uses to make decisions about housing, benefits, or policing, could you? In most places, the answer is no. Public registers of AI systems aim to change that by creating searchable databases of AI tools used by government agencies and, in some proposals, by private organisations operating in regulated sectors. The idea is simple: if an AI system affects your life, you should be able to find out it exists. Several cities and countries have begun experimenting with AI registers. Amsterdam and Helsinki were early movers, publishing details of municipal AI systems including their purpose, data sources, and human oversight arrangements. The EU AI Act requires a public database for high-risk AI systems registered in the EU. The practical challenges are real - defining what counts as an "AI system" for registration purposes, keeping registers current as systems change, and providing enough detail to be useful without overwhelming non-expert readers. For public sector organisations, registering AI systems is becoming a compliance requirement. For private sector organisations, the trend is worth watching: what starts as a public sector obligation often expands.