Well-intentioned is not always well implemented, let alone well communicated. Especially when it comes to new EU-wide regulations: In this case, the European Law on Artificial Intelligence (AI Act). The EU countries worked on the draft law for 5 years, which was finally published in July 2024 and came into force in August. Transitional periods of between 6 and 36 months apply, and the majority of companies must comply with the AI Act from August 2026. The EU wants to give companies time to voluntarily implement the requirements. If the deadlines expire, very harsh penalties threaten: up to 7 percent of global sales or 35 million euros. Significantly more than under the General Data Protection Regulation (GDPR). There is still a lot of educational work to be done, as a Bitkom survey of around 600 companies in Germany shows.
According to Bitkom, 21 percent are currently dealing with the AI Act, only 3 percent have already dealt intensively with the new law. Another 29 percent at least intend to deal with it.
At the same time, however, 16 percent say that they will not concern themselves with the AI Act in the future, and around one in four companies (24 percent) have never heard of the long-discussed EU regulation. In total, that means that 4 out of 10 companies are – as of today – pretty unimpressed by the AI Act. The attitude that ‘we don’t use AI, so a corresponding EU law won’t affect us now or later’ is evidence of widespread ignorance. “AI will find its way into more and more applications and more and more companies. The AI Act is binding for everyone and that also applies if individual employees use AI without the company’s knowledge,” says Bitkom President Ralf Wintergerst. This unregulated use of applications has been known since cloud marketplaces came into existence; keyword: shadow IT.
Great uncertainty about the consequences of the AI Act
According to Bitkom, there is also great uncertainty in the economy about what consequences the AI Act will have. 62 percent of companies say the AI Act will make the development and use of AI legally secure. 53 percent assume that it will increase trust in AI. Conversely, however, 45 percent fear that the AI Act will hinder the development of AI in Europe, and 41 percent believe that it will hinder the use of AI in Europe.
A broad majority of 69 percent of companies believe they need help in dealing with the AI Act. Wintergerst: “We must not repeat the mistakes of the General Data Protection Regulation with the AI Act. The existing and newly created supervisory and market surveillance authorities must not only monitor and sanction, they must above all provide assistance on how artificial intelligence can be developed and used in Germany in a legally compliant manner.”
