Discussion on AI Communication Systems

During a segment on CNN’s “The Lead,” Scott Kupor, the Director of the U.S. Office of Personnel Management and Vice Chair of the AI/SI task force, emphasized the need for developing a notification and communication system regarding rogue AI models. He stated, “we need to figure out action plans and ensure that we strengthen the activities of companies involved. This includes establishing notification mechanisms and contingency plans across various agencies.”

Host Jake Tapper posed a question about the transparency of AI companies, asking if they have any obligation to inform the government following an incident. He noted that much of the knowledge regarding rogue models comes from companies like Anthropic and OpenAI.

Kupor responded that creating an effective notification system is a priority for the task force. He suggested that companies should adopt a reporting model to share information when issues arise. Then, he explained, it’s up to the government to determine the appropriate response: “Is this something for the FAA? We need to communicate with the Department of Transportation.” He expressed his commitment to engaging not only with AI model developers but also with various infrastructure sectors, such as financial and energy companies, where security incidents could have significant impacts. The objective, he pointed out, is to find a reliable notification and communication approach, along with actionable plans, to assure the American people that they can trust the technology and feel safe.

Kupor concluded by reiterating that they will enhance their oversight of company activities, ensure effective notification strategies are in place, and develop contingency plans throughout different agencies. He hopes the government will be well-equipped and informed to handle any emerging scenarios involving AI technologies.