Bitkom on the agreement on the AI Act

Berlin, 9 December 2023 - After intensive marathon negotiations, the European Parliament, the Council of the European Union and the European Commission agreed on a political compromise on the EU AI Act in their trilogue negotiations last night, with the final formal decision to be taken next year. Dr Bernhard Rohleder (CEO, Bitkom) explains:

"The agreement on the AI Act is a political showcase success at the expense of the economy and society. The compromise reached last night overshoots the mark, particularly when it comes to regulating generative AI, and interferes deeply with the technology. The EU is thus tying a regulatory block on companies' legs. There is a great risk that European companies will not be able to keep up with the rapid technological development in future due to impractical plans. There has been an unnecessary departure from the previously intended application- and risk-based approach. The so-called general purpose AI models are regulated as a technology in themselves, regardless of the specific application. The proven and fundamentally welcome instrument of mandatory self-regulation through codes of practice is to be utilised. However, these are only intended for a small proportion of the overall requirements, with the majority being rigidly set out in the law itself. There is therefore a risk that this particularly important field of AI will be restricted by requirements that cannot be adapted to new technological developments. The negotiations on the AI Act itself have already shown how difficult such necessary changes are likely to be.

A fundamental compromise has now been reached on paper on this central and highly controversial issue. The next major challenge will be to translate the nightly agreement into practical rules that create a basis for the responsible use of AI. There is still a risk that we will prevent the use and development of AI from Europe instead of enabling it.

Experience with the General Data Protection Regulation also shows that it is not enough to record the right objectives on paper. The challenge is to strike a real balance in practice between risk management and the promotion of innovation. Europe has the opportunity to play a pioneering role in the ethical and responsible development of AI. But this also requires an unbureaucratic and well-structured implementation of the legal act. To this end, a targeted dialogue between all parties involved must be initiated now. Companies need swift legal certainty and practical support in implementing the AI Act."