The European Union is planning to invest €50 billion in development of AI. The sum was announced by European Commission President Ursula von der Leyen in February. The new InvestAI investment complements a pledge of €150 billion from private sector companies made earlier the same week. Combined, the EU aims to invest €200 billion in AI by 2030.

The EU has also announced that it is cutting regulation of AI, including safeguards for citizens affected by use of AI. The European Commission has proposed dropping the EU AI Liability Directive, which would introduce a new liability regime for AI and reduce the burden of proof on complainants. The Commission dropped the directive following criticism voiced by US Vice President JD Vance, speaking in Paris. Vance argued that the European regulatory regime around AI was discriminatory against US tech companies, and singled out the EU AI Liability Directive especially.

The planned investment of €200 billion includes €20 billion for the construction of four AI gigafactories. These AI gigafactories will be used to train large AI Models. Each data centre is planned to have around 100,000 advanced AI chips, significantly more than the AI factories currently being being built.

The new funding complements EU plans announced in December to build twelve AI factories. Seven have already been announced, and a further five will be announced shortly.

In addition, EU supercomputers will be used in the interim to train large AI models and develop new AI applications.

Plans to improve citizen safety and potential redress from increased uses of AI in society and by businesses have been planned since 2022. The EU AI Liability Directive is seen as a complement to the EU AI Act, and arose from concern about the impact that AI may have on individuals and society.

The Directive has faced internal criticism, particularly from Europe’s centre-right European People Party coalition, which includes Germany’s CDU and CSU parties. Some Commission and Parliament figures believe that the EU AI Law, which is only now entering into force, should be allowed to bed down. Some figures believe that the directive may be unnecessary, and that it would be better to wait to see how the AI Act affects practice and individuals.

The directive has also come under attack from US companies. Vice President Vance’s speech in Paris appears to have prompted the final rejection of the directive, at least for the moment, by the Commission.

However, lawmakers in the European Parliament have rejected the European Commission proposal to drop the AI Liability Directive. Axel Voss, the German MEP responsible for steering the AI Liability directive through the European Parliament, described the decision by the Commission as ‘a strategic mistake’. EP lawmakers could exploit a loophole to keep developing the directive and push it into law. The decision lies in part with the Parliament’s Legal Affairs committee. However, the Parliament’s Internal Market and Consumer Protection Committee, which is also responsible for development of the directive, has already voted to continue working on the directive and push for its adoption.