Papers
In-House Evaluation Is Not Enough: Towards Robust Third-Party Flaw Disclosure for General-Purpose AI
arXiv.org published a research paper examining gaps in how flaws in general-purpose AI systems are reported and evaluated compared to established software security practices. The paper, developed by collaborators across security, machine learning, law, and policy, proposes standardized flaw reporting formats, broad disclosure programs with legal protections for researchers, and improved infrastructure for coordinating reports across AI providers.
