Berlin, 09.02.2022

The German AI Association (KI Bundesverband) expresses its concern with the developments to date with regard to Artificial Intelligence (AI) regulation in the European Union. Since the publication of the EU Commission’s proposal in April 2021, some things have happened, but far from enough.

Currently, the regulatory proposal is at the European Parliament, which clarified the responsibilities only a few weeks ago. In addition to the Committee on Internal Market and Consumer Protection, the Committee on Civil Liberties, Justice and Home Affairs is also responsible for the legislative proposal.

“The European Parliament is indeed aware of the challenges for the European AI ecosystem. At the same time, however, valuable time is being lost to position Europe as a globally competitive location for “AI made in Europe”, explains Nicole Formica-Schiller, board member of the German AI Association, the situation. The fact that different, often competing committees are now responsible for the legislative proposal poses the risk of further delays.

Only last November, the Committee on Artificial Intelligence in the European Parliament published a report that took a closer look at the potential of AI. The European Union is already for a while facing major challenges in this sector and is falling further and further behind in competition with the USA and China but also other global players. Only if the right framework conditions – political, economic and social – are created as quickly as possible, if legal uncertainties are clarified and the digital infrastructure is expanded, will the European AI ecosystem still have a chance against the global competition.

The German AI Association makes it clear that the report contains important points. For example, it said there is a need for clear criteria for high-risk applications, and mandatory conformity assessments would place too much of a burden on the European AI ecosystem. “At least we have the impression, the Parliament puts an end to the rumor that there could be bias-free algorithms or full transparency. This is already a right and important step. In general,
it is advisable for policymakers to work more closely with practitioners, to approve an AI regulation which is usable in everyday work”, Formica-Schiller says.

The German AI Association hopes for a timely concretization of its previous regulatory proposal at the European level in 2022. According to Formica-Schiller, the problems of the General Data Protection Regulation should not be repeated: “Instead of dystopian future scenarios, regulation must focus on the potential of AI. Otherwise, we will see increasing dependencies in the area of AI related to foreign providers as we have already seen in other digital areas.”