The bottom line: The German government is creating a dedicated AI security institute to independently develop its analytical capabilities for large language models.
The German government is planning the establishment of a new institute for security analysis of AI models. This is intended to strengthen expertise in handling large language models such as Claude.
The planned AI security institute is to expand the federal government’s analytical capacities in the assessment of AI systems. The new facility will focus on reviewing and analyzing language models in order to assess their risk potential.
For chief data and chief information officers in the public administration, this means a strengthening of the national governance infrastructure for AI deployments. Such an institute can serve as a reliable source for security assessments and thus form the basis for responsible AI procurement and integration in government agencies.
Federal Minister Wildberger emphasizes the institute’s ambition to provide “world-class experts”. This addresses both regulatory requirements such as the EU AI Act and the practical necessity to independently evaluate AI risks rather than relying exclusively on manufacturer specifications.
Source: www.heise.de · Published 9 June 2026
Lumi AI News — AI-assisted curation pursuant to Art. 50 EU AI Act. Paraphrase and classification by Lumi News Pipeline v1.6.5.