Ahead of tomorrow’s cabinet discussion of the AI Implementation Act, eco – Association of the Internet Industry welcomes the 1:1 implementation of the EU requirements, but warns of practical enforcement problems and unclear responsibilities.
Uniform and practical implementation is crucial
“Artificial intelligence is a central key factor for Germany’s competitiveness and digital sovereignty. It is now crucial that the AI Regulation is implemented not only in a formally correct manner, but also in a uniform, practical and innovation-friendly manner,” explains Oliver Süme, Chair of the Board of eco – Association of the Internet Industry.
eco particularly welcomes the planned role of the German Federal Network Agency (BNetzA) as market surveillance authority and as a central point of contact and coordination. Concentrating expertise at BNetzA can help to support smaller public authorities and strengthen the coherence of supervision. Nevertheless, it remains unclear how a uniform interpretation of the AI Regulation will be ensured in practice. Although a coordination and competence centre is planned, it has no directive powers.
“Without binding cooperation mechanisms, there is a risk of differing interpretations and enforcement practices. This would undermine the desired European coherence and create new legal uncertainties,” says Süme.
Strengthening participation and promoting innovation in a targeted manner
eco explicitly welcomes the obligation enshrined in the draft law to involve civil society, trade unions, business, science and research, as well as the federal states in an appropriate, transparent and regular manner. However, this participation should be formalised to ensure its effectiveness. In eco’s view, the establishment of an advisory board would be appropriate to structurally involve relevant stakeholders, identify implementation challenges at an early stage and support practical solutions.
eco attaches particular importance to the planned measures to promote innovation, especially the AI regulatory sandboxes. However, these must also have sufficient capacity to enable realistic access for small and medium-sized enterprises and start-ups. In addition, at an early stage it should be examined how the real-world laboratories can be meaningfully integrated with other European initiatives, such as the AI Factories or Data Labs from the EU’s AI Continent Action Plan. “Only through close coordination can duplicate structures be avoided and synergies effectively exploited,” emphasises Süme.
Focus on coherence, legal certainty and proportionate enforcement
eco therefore calls for the implementation of the AI Regulation to be consistently geared towards European coherence, legal certainty and the promotion of innovation. This includes the broadest possible centralisation of market supervision, binding coordination mechanisms between the competent public authorities, ensuring that structures are adequately staffed, and creating a proportionate, Europe-wide uniform sanctions framework that addresses risks without deterring SMEs in particular from investing in AI.
