ACCIS response to the European Commission’s public consultation on high-risk AI systems

In our response we explain that credit information suppliers are using AI for a number of different use cases ranging from KYC and anti-fraud processes, data filtering (not high-risk) all the way to producing credit scores (high-risk). At the same time, we underline the fact that the majority of credit scores produced by credit information suppliers are done so using Logistic Regression, which is not an AI system.

This consultation also touches upon important areas such as the interplay between the EU AI Act and the GDPR, as well as between the Act and financial services legislation. ACCIS and its members would appreciate further clarification on the interplay between the AI Act and these other pieces of legislation, not least in terms of how it concerns the use of specific types of data but also on a more general level and how the various requirements can be simplified. In particular, credit scoring is already under strict supervisory scrutiny by sectoral regulation, and at national level.

The Commission is required to provide guidelines specifying the practical requirements of high-risk AI systems and high-risk classification by 2 February 2026. With this consultation, the Commission aimed to receive input from stakeholders on issues that should be clarified in the guidelines.