Claudia Plattner, President of the Federal Office for Information Security (BSI), has strongly criticized leading AI providers for their deficient approach to cybersecurity. Speaking to Spiegel, Plattner pointed out that these manufacturers have paid far too little attention to securing their systems, noting a lack of commitment and seriousness required to prevent these issues. This criticism applies to major providers globally, including those in the United States and China.
Plattner urged these manufacturers to improve the prevention of uncontrolled attacks originating from their systems, insisting that they must stand accountable for any fallout. She stressed that since the products belong to the companies, they must ensure such incidents are prevented in the future; otherwise, serious questions of liability could arise.
With new IT security incidents constantly emerging due to unchecked AI agents, Plattner stated that no one can be satisfied with the current status quo. She argued that regulation is necessary, and it must go beyond self-imposed commitments and mere promises to improve security down the line. The existing political instruments are not sufficient for this purpose. “We must address this technically: What kind of AI is this? Who owns it? Who commissioned it? What is it allowed to do? We need to adapt our rules to this new era and the phenomenon of autonomously operating AI agents.”
Furthermore, she emphasized the requirement for an oversight body dedicated to verifying compliance with these new regulations. Plattner does not anticipate that the IT security problems caused by AI agents-and the human misuse of AI-will quickly subside. She concluded by telling Spiegel that difficult times lie ahead.
