This publication summarizes the key takeaways from workshops overseen by the Center for Security and Emerging Technology (CSET) at Georgetown University and the Program on Geopolitics, Technology, and Governance at the Stanford Cyber Policy Center. The goal of this report is to “provide a high-level discussion of AI vulnerabilities” and “articulate broad recommendations as endorsed by the majority of participants at the workshop.” With the understanding of the urgency to address threats to AI systems, participants outlined key suggestions such as: “1. Extending Traditional Cybersecurity for AI Vulnerabilities. 2. Improving Information Sharing and Organizational Security Mindsets. 3. Clarifying the Legal Status of AI Vulnerabilities. 4. Supporting Effective Research to Improve AI Security.” The workshop noted the challenges of countering AI threats due to their rapid evolution and distinctness from traditional cybersecurity protection mechanisms. As one participant states, “Adding a machine learning model [to a product] is two lines of code; adding defenses can take hundreds.”
