AI action plan vague on protections needed

This month the Commission launched it’s AI Continent Action Plan, promising investments of up to €200 billion, the creation of five AI gigafactories and thirteen AI factories across the EU. This move comes after copious legislative work in the field of artificial intelligence, and will be followed by a apply AI proposal in the coming months.

Looking to boost and leverage the use of AI in specific areas such as the public sector and healthcare, we welcome one of the Commission’s core tenants being the use of trustworthy and human centric models, yet note the lack of detail surrounding a number of key issues.

Commission President von der Leyen’s first mandate was centred on the introduction of the European Green Deal, which has since been scaled back significantly through the new simplification agenda, and is under consistent attack from the far-right and European People’s Party. In line with the new mantra of competitiveness above all else, the environmental impacts of tripling Europe’s data centre capacity by 2035 is a clear afterthought. Only one reference to the need for sustainability is mentioned, with no clear pathway to delivering these lofty ambitions while aligning with a reduction in carbon emissions; “envisages that data centre projects that meet requirements related to resource efficiency, including in energy and water efficiency, circularity and requirements related to innovation will benefit from simplified permitting, while maintaining environmental safeguards and protecting human health”.

As outlined by Eurocadres last year, the use and development of AI systems should not be exempt from sustainability targets and potential sanctions. A report from 2024 cites estimates from the International Energy Agency that on an industry-wide level, energy use from data centres that power AI will double in just the next two years, consuming as much energy as Japan. These data centres and AI systems also use large amounts of water in operations and are often located in areas that already face water shortages.

Transparency, safety for workers and from risks, and accountability where shortcomings are found must be central components of future AI development in an already changing Europe, yet these are lacking from the proposed plan. An omnibus on AI regulation has also been announced for the end of this year, once again underlining the intention to cut back on legislative protections.

The need for training to seize any opportunities AI may present is thankfully noted, yet we again repeat that without a binding right to training, during working hours and at no cost to workers themselves, the skills gaps employers decry will continue to mount.

While there is plenty for business to be happy about, the leveraging of such funds without references to quality jobs, fair and efficient public procurement that shows clear criteria to avoid the misallocation of public funds to companies who circumvent European rules, environmental concerns and reasonable sanctions once again highlights the direction of travel for this Commission college.