Following direct export controls imposed by the US government on AI software, Anthropic has blocked its Fable 5 and Mythos 5 models worldwide. The eco Association has warned of risks to Europe’s cybersecurity: powerful AI models are vital tools for analysing vulnerabilities, detecting attacks more quickly and strengthening IT systems. Europe must build up its own capabilities in developing, operating and auditing such models.

With the AI Implementation Act, the German Bundestag is laying the groundwork for the national implementation of the European AI Regulation. eco Association welcomes the one-to-one transposition of the EU requirements and the central role of the German Federal Network Agency (BNetzA), but still sees a clear need for improvements in terms of coordination and enforceability. Without clear responsibilities and uniform enforcement practices, there is a risk of legal uncertainty, additional bureaucracy and barriers to investment for companies. It's now crucial that AI regulation is implemented in a practical, innovation-friendly and consistent manner across the country.