Co-rapporteurs working on the AI file in the Parliament have put forward the idea of creating an AI Office to oversee the enforcement of the AI Act and mediate when it comes to cross-border cases.
As part of new compromise amendments circulated last week and seen by EURACTIV, lawmakers proposed a centralised approach that would replace the original AI Board with a new body, just shy of a new EU agency, with its own legal basis, funding and staff to be tasked with supporting enforcement and resolving disputes. While there is broad consensus on the need for effective enforcement, the idea of creating an AI enforcement agency of any kind has received some pushback from conservative MEPS (and member states) in the past.
Related: EC Proposes “EU Data Act”, Forcing Manufacturers To Share Data
Also on the menu of the shadow meeting on Wednesday was the definition of Artificial Intelligence. The definition was removed from Annex I, as the co-rapporteurs consider it such a critical aspect of the legislation that it cannot be changed via secondary legislation later.
The new wording is more outcome-oriented than focused on the process. In particular, AI would have to satisfy three cumulative conditions: receive machine or human-based data; infer how to achieve a given set of goals using learning, reasoning or modelling; generate outputs in the form of content, predictions, recommendations or decisions influencing the real or virtual environment it interacts with. The discussions on this fundamental point are still at a very early stage.
