Skip to main content

Security

AI Risk Profiles: A Standards Proposal for Pre-Deployment AI Risk Disclosures

Sherman & Eisenberg (2023)

Category
Risk Domain

Vulnerabilities that can be exploited in AI systems, software development toolchains, and hardware, resulting in unauthorized access, data and privacy breaches, or system manipulation causing unsafe outputs or behavior.

"Encompasses vulnerabilities in AI systems that compromise their integrity, availability, or confidentiality. Security breaches could result in significant harm, ranging from flawed decision-making to data leaks. Of special concern is leakage of AI model weights, which could exacerbate other risk areas."(p. 23048)

Other risks from Sherman & Eisenberg (2023) (8)