Politicians from the Union (CDU), the Social Democrats (SPD), and the Greens are raising alarms over increasingly capable, freely accessible AI models. Marc Henrichmann, the chairman of the Bundestag’s Intelligence Services Control Committee and a CDU member, stated in the “Handelsblatt” that powerful, highly developed open-source AI models could fundamentally alter the security situation. According to Henrichmann, publishing model weights or other essential components significantly speeds up innovation cycles because developers can modify, improve, and further distribute the systems. However, this simultaneously lowers the barrier for cybercriminals and state actors to employ these systems for automating attacks, running influence campaigns, or developing new malicious software.
Johannes Schätzl, the digital policy spokesperson for the SPD parliamentary group, anticipates a significant rise in security risks due to the deployment of powerful open AI models. He identified cybercriminals as the primary danger, stressing that they could operate open models locally, modify them, and use them without the oversight of the original provider. Schätzl warned that this could reduce the hurdles for launching cyberattacks, allowing known attack methods to be utilized faster and on a larger scale.
Konstantin von Notz, the vice-chairman of the Intelligence Services Control Committee and a Green member, expects open-source AI to cause massive disruption within IT security. While he acknowledged that open models strengthen transparency and digital sovereignty, he pointed out that they also provide criminals and other hostile actors with easy access to potentially very powerful tools. Von Notz concluded that the state of IT security will fundamentally change as a result, requiring decisive counter-measures from the federal government, with a particular focus on rapidly closing security vulnerabilities.
