Following its AI Act coming into full force this August, the EU is taking consultations to develop guidance for compliance.
The European Commission is seeking input from relevant parties and EU citizens on two different matters, including how the law defines what an AI system is, and what AI applications should be banned.
The first consultation aims to create a more precise definition of what an AI system is, seeking opinions from businesses, academic and civil societies, as well as the AI industry itself.
Respondents are tasked to rate the importance of different facets of AI’s definition as it currently stands in the AI Act, from ‘a machine based system’ and ‘predictions, content, recommendations, or decisions,’ to ‘designed to operate with varying levels of autonomy’ and ‘for explicit or implicit objectives.’
It also askes respondents to provide examples of software systems or programmes that do not fall under the scope of the AI system definition provided in Article 3(1) of the AI Act.
The second portion of the consultation deals with instances where AI would be banned if the use case is deemed to be harmful.
Respondents are asked if certain stipulations or reasons for the bans require clarification in order to make the law as comprehensive and understandable as possible.
Further, it asks respondents for any concrete examples of existent AI systems that fall under these restrictions.
Recommended reading
- The EU Agreed on an AI Act, What Now?
- EU AI Act “Riddled” with Loopholes
- Where Will Companies Look to For AI Regulation Guidance?
AI systems that use facial recognition, emotional recognition, or decision making based on scoring face certain restrictions, and AI systems that have the potential to be misused for malicious activity also face bans.
The consultation, which provides information about the current Act as it stands, will run until 11 December 2024, but the Commission is urging for early responses.
It expects to offer an update on banned uses and definitions of AI systems by early 2025.





