Parliament mulls monitoring group for AI Act implemention

The European Parliament is discussing the establishment of a monitoring group for the implementation of the Artificial Intelligence (AI) Act, which comes into force in August, a leading EU lawmaker for the file told Euractiv.

MEPs have expressed concerns over the direction that the European Commission, tasked with overseeing implementation, appears to be taking on key tasks.

“As the outgoing rapporteurs and shadow rapporteurs on the AI Act, we had discussions with the Commission and will soon formalise a monitoring group from the Parliament,” center-left MEP Brando Benifei (S&D) and former rapporteur on the EU’s landmark regulation on AI, told Euractiv on Thursday (18 July).

“I will for sure have a leading role [in this oversight body],” he added.

The matter will be discussed among coordinators of the Parliament’s committee on Internal Market and Consumer Protection (IMCO), likely next week, two people familiar with the matter told Euractiv.

The Parliament already has structures within committees to monitor the implementation of the Digital Services and Digital Markets Acts, two landmark EU regulations on content moderation and digital competition.

The AI Act’s implementation falls under the jurisdiction of national regulatory authorities, with the Commission’s AI Office  tasked with key coordinating responsibilities.

The Commission has already started building the AI Office and the Act will come into force on 1 August 2024, with a staggered timeline for when provisions are applicable.

Benifei cautioned that there are risks of the AI Act losing some of its intended impact if the Commission fails to implement it properly, particularly regarding transparency provisions like watermarking and copyright protection.

He said the Commission’s AI Office needs to operationalise fundamental rights impact assessments with clear and practical templates to address risks to citizen’s fundamental rights.

Red flags

Three MEPs inquired about the lack of transparency in the Commission’s hiring strategy for the new AI Office, which reportedly includes the other rapporteur of the Act, former Renew MEP Dragoș Tudorache.

Benifei declined to comment on the Commission’s recruitment processes, noting that they have only just begun.

Some MEPs, including Benifei, “strongly urged” the Commission in a letter dated 3 July to involve civil society in developing codes of practice for general-purpose AI, as opposed to letting the companies write the codes themselves.

Euractiv has reported that the Commission is ploughing ahead with hiring consultancies to manage the drafting process.

Powerful general-purpose AIs like ChatGPT present some of the most concerning and controversial aspects among regulators because of their wide-ranging abilities.

The group of MEPs that sent the letter included communists, socialists, greens, liberals, Christian-democrats, and a member of hard-right ECR group.

“We received reassurances from [Single Market Commissioner Thierry] Breton, who informed us in a written statement that civil society and independent experts will be fully involved in drafting the codes of practice for the most powerful models,” Benifei said.

Benifei said that while Big Tech should contribute as stakeholders, they should not lead the drafting of these codes of practice.

The Italian MEP wants to push for having independent experts chairing the working groups inside the AI Office to curb the influence of large companies, he said.

“Without harmonised standards, we will not have an AI Act,” Benifei warned.

“The urgent task now is to implement digital regulations, build a strong AI Office that works efficiently with national authorities, and ensure supervision by the European Parliament,” he said.

[Edited by Zoran Radosavljevic]

Read more with Euractiv