We use cookies to personalise the website and offer you the greatest added value. They are, among other purposes, used to analyse visitor usage in order to improve the website for you. By using this website, you agree to their use. Further information can be found in our data privacy statement.



Update on the EU AI Act: German Draft AI Market Surveillance and Innovation Promotion Act (KI-MIG)

​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​published on 24 February 2026​ | reading time approx. 2 minutes


Since 1 August 2024, Regulation (EU) 2024/1689 (“EU AI Act”) has established a directly applicable legal framework for AI systems across the European Union. The Regulation obliges Member States in particular to designate competent authorities, ensure market surveillance, and cooperate with European bodies.

In Germany, a specific national enforcement framework had not yet been adopted. On 11 February 2026, the Federal Government therefore approved a draft AI Market Surveillance and Innovation Promotion Act (KI-MIG) (Bundesrat document 97/26). 

The Act does not create substantive AI law of its own; rather, it regulates the organizational implementation and allocation of responsibilities for enforcing the EU AI Act in Germany.

Central coordination structure​

The core element of the draft is the establishment of a Coordination and Competence Centre for Artificial Intelligence (KoKIVO) at the Federal Network Agency (Bundesnetzagentur). Its functions include:
  • coordinating cooperation among competent authorities;
  • consolidating horizontal legal and interpretative questions;
  • providing expert support in complex assessments of AI systems;
  • acting as the national contact point vis-à-vis European bodies.

The material competences of sector-specific and market surveillance authorities (e.g., product safety, medical devices, or financial supervision) remain unaffected. KoKIVO therefore does not replace specialised regulators but serves coordination purposes.

AI regulatory sandboxes​​

The EU AI Act requires Member States to establish AI regulatory sandboxes in which AI systems can be developed and tested under the supervision of competent authorities. The objective is to clarify regulatory requirements at an early stage and to prepare the later conformity assessment.

Participation in a sandbox does not constitute an authorization or a legal privilege. In particular, it does not replace conformity assessment and does not provide liability protection.

Cooperation with data protection authorities​​

AI systems frequently process personal data, leading to overlaps with data protection supervision. The draft law provides for organizational coordination and the forwarding of inquiries to the competent authorities.

The competences of data protection authorities remain unchanged. Accordingly, the Act aims to im-prove administrative coordination rather than to establish a single supervisory authority.

Legislative process​

The draft has first been submitted to the Bundesrat. After completion of the consultation procedure, it may be introduced into the Bundestag.​

Tech & Data Bites

author

Contact Person Picture

Johannes Marco Holz, LL.M.

Attorney at law (Germany)

Partner

+49 911 9193 1511

Send inquiry

Rödl Germany

Discover more about our offices in Germany​​. Re​​a​d more »
Skip Ribbon Commands
Skip to main content
Deutschland Weltweit Search Menu