Cities around the world have been adopting algorithmic systems to address the most urgent urban challenges including sustainable urban mobility, energy efficiency, housing and healthcare.
AI offers enormous opportunities for cities to promote digitalised and more efficient service provision. At the same time, however, AI solutions also bring about risks associated to the automation of discrimination, the deepening of existing inequalities, lack of transparency or the threat to privacy.
To minimise these risks and foster transparency and communication with citizens, solutions such as AI registers offer opportunities to document decisions and assumptions made by algorithms in cities in a standardised way.
Based on the good practices from the cities of Helsinki and Amsterdam developing and implementing algorithm registers, and in collaboration with nine European Cities -Barcelona, Bologna, Brussels, Eindhoven, Mannheim, Rotterdam and Sofia-, the Eurocities Digital Forum Lab has developed a Data Schema for algorithm registers.
Freely accessible and open source, the data schema for algorithm registers sets common guidelines for cities on the information to be collected and published on algorithms and their use.
The AI Register offers a standardised, searchable and archivable way to document the decisions and assumptions that were made using AI systems. The information included can help citizens to get an overview of algorithms used by governments, improving public scrutiny and measuring the impact on their lives.AI registers also offer a tool to local governments to manage AI governance, promoting more transparent and responsible AI.
The Cities Coalition for Digital Right collaborates with Eurocities to improve digital governance of local administrations and to mitigate the impact of the use of algorithms on human rights.
If your city is interested in scaling the data schema or you would like to tell us about your experience, make sure to contact email@example.com
The full data schema can be viewed and downloaded here: https://www.algorithmregister.org/