APPLiA's Feedback on the Draft Commission Guidelines on the Classification of High-Risk AI Systems

APPLiA supports the Commission's objective of ensuring a consistent application of the AI Act, in particular the objective of supporting businesses with guidance documents that support the application and understanding of the classification logic within the AI Act as well as requirements for providers or deployers.

However, the draft guidelines in its current form do not sufficiently resolve key concerns of manufacturers operating under Union harmonisation legislation listed in Annex I AI Act. While several clarifications are helpful, the current text continues to create legal uncertainty for manufacturers that voluntarily introduce AI-enabled safety improvements into already compliant and safe products.

The draft guidelines risk creating a situation where manufacturers are discouraged from developing and deploying additional AI-based safety measures because doing so may trigger high-risk classification and substantial compliance obligations, despite the product remaining safe without the AI functionality.

APPLiA therefore recommends targeted amendments focusing on:

• Correcting the interpretation of the wording “required to undergo third-party conformity assessment”  in Article 6, 1 (b)
• Clarifying the interpretation of "safety component" as the key concept for high-risk classification in Article 6, 1 (a)
• Clearly differentiating between critical safety functions and minor, ancillary safety layers that are only installed additionally to already fully compliant products (in line with Annex I legislation).