The EU’s Ambitious AI Action Plan Needs to Translate into Tangible Solutions and Data Use in the Industry
The AI Continent Action Plan, published by the European Union on 9 April, sets out an ambitious agenda. Encouragingly, it also acknowledges the real-world challenges companies face in leveraging data and artificial intelligence. It is now more important than ever to ensure the implementation of innovations to strengthen the competitiveness of European industries.
It is welcomed that the Commission is committed to the fundamental prerequisites for AI development. There are large figures – €200 billion for AI development and €20 billion for large-scale supercomputers (so-called AI Gigafactories). However, the majority of funding for these AI Gigafactories is expected to come from member states and businesses.
One positive aspect is that investments in AI infrastructure may spill over into other areas of critical technology. These investments could, for example, attract much-needed semiconductor manufacturing projects to Europe.
Perhaps the most striking figure in the action plan is 13.5% – the proportion of European companies currently making use of AI in their operations. This number needs to rise significantly if Europe is serious about improving its competitiveness. To this end, the Commission is preparing an Apply AI strategy, which must in our view focus on two priorities: co-developing sector-specific AI models and solutions and promoting business-driven AI networks. A good example of the latter is the AI Finland network, initiated by the Technology Industries of Finland.
Using AI requires data – and unfortunately, the current state of data utilisation in Europe is not much better than the adoption of AI itself. While the plan’s aim to pave the way for data use in AI development is laudable, it must be placed in a broader context. In practice, most data sets include personal data, and this makes their use in industry difficult. Standardised and scalable solutions are yet to be developed for industry, largely due to heavy and complex personal data regulation and the often-cryptic guidance issued by authorities.
The Commission’s forthcoming European Data Union strategy should therefore set an ambitious goal: clear and scalable mechanisms for making data sets – even those containing personal data – readily usable in industry.
Industry needs clear practices to support data-driven growth – not a maze of overlapping actions. For instance, the Commission’s proposal to establish new Data Labs in conjunction with AI Factory supercomputers is ill-timed. Rather than launching new initiatives, the focus should be on developing existing ones. Both AI Factories – such as Finland’s LUMI AI Factory – and the Data Spaces established in the EU’s Data Act are still in their infancy – these initiatives should be pushed forward, together with the industry. Overlapping and fragmented structures must be avoided.
Further information
