The new provisions of the AI Act include both a ban on certain AI practices and requirements for employees' qualifications in dealing with artificial intelligence (AI). These include social scoring systems, manipulative AI techniques and emotion recognition in the workplace.

The digital association Bitkom expresses concerns about the practical implementation and criticizes insufficient preparation on the part of politicians. Susanne Dehmel, Member of the Bitkom Executive Board: »The AI Act was supposed to provide legal certainty for artificial intelligence in Europe — but the opposite now seems to be the case. With further provisions of the European AI regulation coming into effect on Sunday, it remains unclear which applications will actually fall under the legal prohibitions. The government has set high requirements and tight deadlines for companies regarding the AI Act, but has failed to do its own homework. The risk is borne by companies that develop or deploy AI. While the U.S. is investing hundreds of billions in expanding AI and China is releasing highly advanced language models, we are throwing obstacles in the way of AI companies in Germany and Europe.

Moreover, the definition of prohibited practices does not solely capture clearly problematic applications. Only today, two days before the regulations take effect, is the EU Commission planning to release urgently needed guidelines on the definition of AI and prohibited practices. Companies whose systems fall into numerous difficult-to-categorize gray areas will find out today, Friday, whether their systems may remain on the market or must be withdrawn over the weekend.

There is also uncertainty regarding the so-called AI competence requirements, which affect virtually all companies that develop or use AI. All companies using AI must ensure that the relevant employees possess "an adequate level of AI competence." When this obligation is considered fulfilled remains unclear, partly because no regulatory authority has yet been established in Germany to provide guidance to businesses. European authorities have also failed to issue specific clarifications.

However, companies should not let this stop their AI efforts. Instead, they should take the regulation as an opportunity to further train their employees in handling AI and prepare them for developments beyond the AI Act.

We need more German and European AI. This requires not only funding and increased support for European AI companies but also a truly innovation-friendly regulatory framework — not an excess of vague regulations but rather a minimum of binding and clear guidelines for practical application. It's high time we stopped viewing artificial intelligence primarily as a threat and instead recognized it as an opportunity.«