NIST Seeks Comment on AI Safety and Detection Tools

The public can comment through June 2 on the National Institute of Standards and Technology’s continued efforts to improve the safety and trustworthiness of AI systems, Commerce Secretary Gina Raimondo announced Monday. NIST issued four draft documents in support of President Joe Biden’s executive order on AI. The first draft document seeks comment on mitigating risks related to generative AI. This includes AI tools for accessing chemical, biological and nuclear weapons. The agency is interested in risks associated with cybersecurity attacks and AI-driven hate speech. NIST is gathering feedback for a second draft document analyzing how data is used to train AI systems, particularly when technology is used for malicious purposes such as the perpetuation of human bias. The agency is studying technical approaches for identifying AI-created fake content. Detection methods include digital watermarking and metadata recording. A fourth NIST document outlines approaches for global development and implementation of AI standards for detecting AI content. NIST is “working hard to research and develop the guidance needed to safely harness the potential of AI, while minimizing the risks associated with it,” said Raimondo.