Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
A 40-minute Black Hat conference talk examines the legal implications of prompt injection attacks on Large Language Models (LLMs) through the lens of U.S. cybersecurity law. Explore how the Computer Fraud and Abuse Act (CFAA) applies to different types of prompt injection, which ranks at the top of OWASP Top 10 for LLMs. Learn about the potential legal risks faced by AI security researchers and red teamers when testing LLM vulnerabilities, highlighted by recent cases like OpenAI's lawsuit against NYTimes over alleged "deceptive prompts." Understand why the Supreme Court's interpretation of the CFAA presents challenges when applied to LLM security testing, and discover why legal safe harbors may be necessary to protect legitimate AI security research. Delivered by experts from Harvard Law School, Osgoode Hall Law School, and UC Berkeley, gain valuable insights into the intersection of adversarial machine learning and cybersecurity law.
