With the AI Application Act, the Federal Network Agency becomes the central coordination and market surveillance authority for artificial intelligence. In the future, it will be the first point of contact for companies and citizens on issues of AI compliance – outside of already regulated sectors such as banking, medical devices, or insurance. In these areas, the specialist authorities remain responsible, while the Federal Network Agency coordinates the overall framework.

From August 2, 2026, the authority can take action on its own initiative, based on complaints, or as part of random checks. Its powers include initiating examination procedures, requesting technical documentation and conformity certificates, and conducting its own tests. If deficiencies are identified, it can demand improvements, order recalls from the market, or issue sales bans.

In addition, the Federal Network Agency will be expanded as an AI competence center. Planned are an AI service desk as a low-threshold point of contact for SMEs and start-ups, a real-world laboratory for legally compliant testing of new AI applications, and targeted networking, support, and training offers regarding the requirements of the AI Regulation.

With the deadline of August 2, 2026, central transparency obligations will also come into force. Companies must label AI chatbots and voice systems as such upon first interaction, disclose emotion recognition and biometric categorization, and identify deepfakes and AI-generated content in public communications, advertising, or political contexts as such.

The fine framework is based on the EU AI Regulation and is nationally designed in the KI-MIG. Fines of up to 15 million euros or three percent of global annual turnover are possible for violations of transparency obligations, and up to 35 million euros or seven percent of turnover for prohibited AI practices. Further fine offenses are provided for certain national violations. The background is the EU AI Regulation, which has been gradually applicable since 2024. With the KI-MIG, Germany is now creating the national enforcement infrastructure. The Federal Network Agency will thus become the central AI authority, which will also focus more strongly on promoting innovation.