European Court of Auditors: Commission failed to effectively coordinate boost in AI investment

The European Commission failed to effectively coordinate a planned boost in investments in artificial intelligence (AI) on par with global leaders between 2018 and 2022, said the European Court of Auditors (ECA) in a report published on Wednesday (29 May).

The Commission declared AI “one of the most strategic technologies of the 21st century” in 2018 when it announced an initiative to boost investment in the technology and increase its uptake among the private and public sectors.

It released two plans to narrow the gap with competitors, one in 2018 and another in 2021. The court said the difference between US and EU investment had doubled by 2020, citing a 2022 technical report by the Commission’s Joint Research Center.

Unfortunately, the gap has grown even more since then. In 2023, the US saw €62.5 billion in private investment, seven times more than the EU and UK’s €9 billion, the European Parliament said in a March 2024 research report citing another report by Stanford University. Stanford found that private investment in AI in the bloc fell by 44% from 2022 to 2023.

The ECA interviewed relevant officers at the Commission and the European Investment Fund and 23 national authorities of member states, 20 of them responsible for implementing AI policy. The court also reviewed Commission documents and looked into 10 AI projects funded by Horizon 2020, Europe’s flagship science research initiative.

The ECA said the targets set were too vague and didn’t assign research priorities or member states’ specific goals to guide investments. Neither were the targets updated as the investment gap with the US grew.

The court said coordination with national authorities was also ineffective, and only one expert group per member state was tasked with coordinating plans.

The EU executive body also failed to comprehensively review, as initially planned, how the 2018 and 2021 AI plans were implemented in member states.

“The Commission has already taken several steps and is very actively engaged in the preparation of the implementation” of the AI Act, including setting up the AI Office, an official for the executive said in response to the ECA report.

The official also drew attention to new measures to boost investment in AI: An innovation package aimed to create “an additional overall public and private investment of around €4 billion until 2027” in generative AI was announced in January 2024.

The Commission also has “made Europe’s supercomputers available to innovative European AI start-ups to train their trustworthy AI models” after Ursula von der Leyen’s 2023 State of the Union speech. Horizon and Digital Europe invest €1 billion annually in AI research, and the Commission plans to attract “more than €20 billion in annual investment” in the EU over the next decade.

According to the initiative’s website, the Commission plans to invest €2.1 billion in AI between 2021 and 2027 under its Digital Europe programme.

Going forward, the court recommends reinforcing the planning of how the EU and member states can reach investment goals and establishing a monitoring framework to assess progress.

Implementing regulations such as the Data Governance Act, which sets out rules for a single data market, and the AI Act, which governs the use of the technology based on risk, is also important, said the court.

[Edited by Rajnish Singh/Alice Taylor]

Read more with Euractiv

Picking up the AI Liability Directive after the tech policy spree

The next European Parliament will have to determine whether a niche piece of liability rules related to Artificial Intelligence (AI) can be useful after the previous parliament's legislative spree.