AI Development Practices Act of 2024

This bill directs the National Institute of Standards and Technology (NIST) to catalog and evaluate emerging practices and norms for communicating characteristics of artificial intelligence (AI) systems, such as transparency, robustness, security, and safety, and to develop voluntary guidance for AI development practices.

Sec. 2, NIST research on development best practices: This section amends the NIST Act to add a new subsection requiring the NIST Director to develop and periodically update voluntary guidance for AI development, release, and assessment practices. The guidance must define methods for setting risk tolerances based on security, economic, and social risks; categorize practices for communicating AI characteristics like robustness, fairness, and privacy; and provide recommendations for using those practices. The Director must update the existing AI risk management framework, base guidance on international standards, avoid prescribing specific technology products, and collaborate with public and private stakeholders. Within 18 months of enactment, the Director must brief Congress on new materials produced and policy recommendations to improve private-sector and federal coordination. The section defines "artificial intelligence red teaming" as structured adversarial testing to find flaws and vulnerabilities in AI systems, and defines "artificial intelligence system" by reference to existing law.