ALLAI’s Policy Analysis on the Council General Approach for the AIA

ALLAI offers a policy analysis on the recently adopted Council General Approach for the AIA

During the December meeting of the Telecom Council, the Member States of the EU adopted their ‘general approach’ regarding the proposal for the Regulation on Artificial Intelligence (AIA). This means that we are roughly half way through the legislative process. The European Parliament will likely adopt its position in March of this year, after which the negotiations between the European Commission, the Council and the EP (Trilogue) will start. What stood out at the Council meeting though, was the fact that a notable number of Member States seemed to accept the approach merely ‘in the spirit of compromise’ signalling ‘room for improvement’ on several points.

In this paper, ALLAI provides an initial analysis on the main elements of the Council’s ‘general approach’.

The paper covers topics such as:

  • Proposal for a new AI-definition
  • The amendments towards stronger bans (covering the exploitation of vulnerabilities, social scoring, and biometrics)
  • Blanket exclusions (i.e. Military, defence and national security, Research and development, Collaboration with third country public authorities and international organisations)
  • General Purpose AI (GPAI))
  • Fundamental rights impact assessment
  • Extra horizontal layer for high-risk AI waters down the AIA
  • ANNEX III deletions
  • Requirements for high-risk AI

The paper expresses our support and concerns towards some of the amendments made, and provides recommendations in response.