5.5. Ethical principles and the algorithm register

The City takes ethical principles into consideration in all utilisation of artificial intelligence and advanced analytics. The City will develop a framework for the responsible utilisation of AI, i.e. a metadata model[1], in collaboration with Finnish and international partners (including Sitra, the Ministry of Justice, Saidot, Espoo and Amsterdam). The metadata model will collect the different elements with the help of which the operating principles of algorithms will be made transparent and understandable. The AI algorithms used in the City’s digital services will be collected in a shared algorithm register, the purpose of which is to document and openly communicate how the models used by the City are built and what data they are based on. This will help the City’s customers to understand the decision logic behind services.

The European Commission’s artificial intelligence strategy[2] also strongly emphasises the transparency and explainability of AI. Europeans must be able to trust the recommendations provided and decisions made by AI. The Commission calls for new, significant investments in data and AI, but also new rules and regulations, especially for so-called high-risk AI applications. As examples of sectors where AI applications can present significant risks, the Commission lists healthcare, police forces and transport. The Commission’s white paper is currently undergoing public consultation and will most likely result in the issuing of an EU artificial intelligence directive in the coming years.

[1] https://www.saidot.ai/resources

[2] European Commission (2020): Artificial Intelligence - A European approach to excellence and trust. White Paper.