Who enforces Article 50 in Germany?

Germany's answer is further along than most Member States'. The national AI Act implementation law, the KI-Marktüberwachungs- und Innovationsförderungsgesetz (KI-MIG), was approved by the federal cabinet on February 11, 2026, and designates the Bundesnetzagentur (the Federal Network Agency, which already regulates telecoms, post, and energy) as the central market surveillance authority for the AI Act. As of mid-2026 the final legislative steps were still completing, so treat the precise contours as settling rather than settled.

For an SME selling into Germany, the practical meaning: transparency-duty complaints (an undisclosed chatbot on a German-market site, unlabelled realistic AI media) would land with the Bundesnetzagentur once the structure is fully operational. The agency is an experienced, process-driven regulator; the reasonable expectation for early Article 50 enforcement is what that kind of authority does everywhere: complaint-driven checks of easily verifiable facts, and corrective orders before fines.

The duties themselves don't depend on any of this German machinery. Article 50 applies from August 2, 2026 as directly applicable EU law. The KI-MIG organises who supervises in Germany, not whether the rules bind you. A German customer, competitor, or journalist can observe a missing chatbot disclosure regardless of the statute book's progress.

Penalties in Germany follow the EU-level scale: for SMEs, the ceiling is the lower of €15M or 3% of worldwide turnover (Article 99(6)), not the higher, as press summaries usually quote, and fines are discretionary against the Act's listed factors.

One Germany-specific practical note: German-market users should get the disclosure in German ("KI-Assistent," with the harmonised "KI" label for content), and German business culture leans on documentation, so the dated-evidence habit (screenshots, configs, review logs) earns its keep twice here.