AI and operational powers Hain questions the chiefs

Jul 1

Hain’s final-round slot first asks all three chiefs about AI’s risks and opportunities. Jäger says mass data requires automated analysis and systems face a race against intrusion; Rosenberg describes deployed AI-supported examination of public social-media profiles for constitutional-loyalty screening and a planned Hamburg-university project for better use and protection. Selen cites an unspecified research-lab agent incident as a warning and requests sustained investment and technology-open legal powers; no independently identified lab event or funding award is established. Hain then asks for concrete examples of what the proposed BfV and BND reforms would permit. Selen answers with access to attack plans on devices, intervention against imminent threats and control of an attack instrument, all conditional on legislative approval, close oversight and restrictions to the most serious state and international-terrorism cases. The BND does not deliver a separate answer to that final question before the chair moves on.

The assisted constitutional-loyalty screening provisions commence on July 1, 2026. Rosenberg describes checks before military entry replacing earlier prospective-soldier screening, using partial automation and AI for faster decisions while distinguishing them from full security clearances. In the AI question she explains examining public social-media profiles and producing relevant applicant profiles more quickly. These are her descriptions of the process and benefits; the entire Bundeswehr Protection Act did not first commence in July.