evidoria

← Back to browse

Good practice Imported

Germany's Marktplatz der KI-Möglichkeiten — A Voluntary Federal AI Register With Documented Gaps

Germany · Berlin · See the Germany profile · See the Berlin profile

Evidence: Observational / pre–post Top 83% 40/100 · Ask Evidence Copilot about this practice

Germany's federal AI register, the Marktplatz der KI-Möglichkeiten, lists roughly 240 government AI applications, but entries are voluntary and over a third omit required risk assessments — a cautionary case en route to a mandatory nationwide rollout by 2027.

240 entries (approx.)
AI applications listed in the federal register (December 2025)
more than one third
Listed applications lacking required risk assessments
Germany's Marktplatz der KI-Möglichkeiten — A Voluntary Federal AI Register With Documented Gaps

Details

Maturity
Scaling
Promoter
Bundesministerium für Digitales und Staatsmodernisierung (BMDS) / IT-Planungsrat
Period
2025–2027 (rollout ongoing)
Region (NUTS)
DE300
Keywords
digital government, AI transparency, algorithm accountability, public administration

Context

Germany's Federal Ministry of the Interior (now the Bundesministerium für Digitales und Staatsmodernisierung) launched the Marktplatz der KI-Möglichkeiten (MaKI) in early 2025 as a voluntary public register where government agencies can list their AI applications.

Objectives

The register aims to give citizens and oversight bodies visibility into how AI is used across public administration, ahead of a mandatory, multi-level rollout.

Activities

Agencies self-report AI applications to the register on a voluntary basis. The IT-Planungsrat, the joint federal/state digital-government body, began expanding the register to state and local administrations in December 2025, with full institutional ownership due to transfer to the IT-Planungsrat by early 2027.

Results

As of December 2025 the register held roughly 240 entries at federal level. Independent analysis by netzpolitik.org and AlgorithmWatch found more than a third of listed applications lack the risk assessments required under the EU AI Act, data is published in an unwieldy spreadsheet format, and information on licensing, partners and legal basis is frequently missing.

Conclusions

MaKI is a genuine, operating register at meaningful scale, but its documented gaps make it a cautionary example of what voluntary, self-reported algorithmic transparency delivers in practice; legal scholars have argued it should become mandatory, citing the Netherlands' Algoritmeregister as a stronger model.

Implementation
Indicative cost
Medium (€50k–€500k)
Time to results
Medium (1–3 years)
Staffing & skills
Bundesministerium für Digitales und Staatsmodernisierung (BMDS), IT-Planungsrat (joint federal/state digital-government body)

Conditions for success

  • Standardised reporting format
  • Mandatory (rather than voluntary) participation
  • Risk-assessment data required under the EU AI Act

Common failure modes

  • Voluntary participation lets agencies choose not to list applications
  • More than a third of listed applications omit required risk assessments
  • Data published in an unwieldy spreadsheet format with missing licensing/partner/legal-basis information
  • Implemented as one of at least five fragmented state/federal initiatives

Where it fits

Governance type
federal ministry transitioning to joint federal/state body
Scale
national, expanding to state and local levels

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful