Report on Public Feedback for AI definitions and prohibited practices published

The European Commission has released a report, prepared by the Centre for European Policy Studies (CEPS) for the EU AI Office, which analyzes feedback from two public consultations on the AI Act. These consultations, conducted in 2024, focused on the definition of AI systems and the list of prohibited AI practices, both of which have been in effect since 2 February 2025. Industry stakeholders dominated the consultation responses, accounting for 47.2% of nearly 400 replies, while citizen participation was notably low at 5.74%. Respondents emphasized the need for clearer definitions of terms such as “adaptiveness” and “autonomy” to avoid unintentionally regulating conventional software.

The report organizes responses to 88 consultation questions into nine key sections, providing a detailed overview of stakeholder concerns. Notably, prohibited practices such as emotion recognition, social scoring, and real-time biometric identification generated significant concern among respondents. Stakeholders requested concrete examples to clarify what is and is not prohibited under the AI Act.

To address these concerns, the Commission has issued non-binding guidelines aimed at clarifying both the definition of AI systems and the scope of prohibited practices. These guidelines provide legal clarifications and practical examples, particularly in areas such as harmful manipulation, social scoring, and real-time remote biometric identification, to aid providers and stakeholders in complying with the AI Act.

The guidelines are expected to evolve as practical experience accumulates and new use cases and questions arise. By offering these clarifications, the Commission seeks to facilitate the effective and consistent application of the AI Act across the EU digital market.