Smashing the ML Stack for Fun and Lawsuits

Explore the legal risks and ethical considerations of adversarial machine learning research in this Black Hat conference talk. Delve into the potential legal consequences researchers face when targeting commercial ML systems from major tech companies. Analyze how existing laws apply to the testing of deployed ML systems, and examine the expectations of vendors regarding system usage. Learn about various attack vectors like evasion, poisoning, and model inversion. Gain valuable insights into relevant legal frameworks, including contracts, the Computer Fraud and Abuse Act, and Section 1201. Conclude with high-level takeaways to navigate the complex intersection of ML security research and legal compliance.