What concrete transparency and information from algorithms and artificial intelligence is due

Abstract

The study details what data, information or knowledge should be provided on AI systems. It is mentioned whether these are present or future regulatory authority obligations, or if they are best practices found in the UK, the Netherlands or France. Thus, information must be provided on the existence of the system, active dissemination obligations and future public algorithm registers; information on who are the algorithm registries; information on who the AI system providers are and who the users are, the data protection officers, why the system protection officers, why the system exists and its specific purposes. Also information about the training data or input data, as well as the type and technical specifications of the AI system, including its code. The information is specified in accordance with Article 22 RGPD. Likewise, information on the guarantees of the IA system, such as impact studies, risk mitigations or human supervision.