The National Institute of Standards and Technology (NIST) has released a draft outline for a proposed AI standards document focused on testing, evaluation, verification, and validation (TEVV) of artificial intelligence systems. The outline, published as part of NIST’s AI Standards “Zero Drafts” pilot, is now open for public comment through September 12, 2025.

The draft aims to provide a flexible, overarching framework that guides practitioners in developing fit-for-purpose TEVV processes tailored to the unique risks and use cases of specific AI systems. It is designed to complement and integrate with existing international standards, such as those under development by ISO/IEC, while supporting broader AI safety and assurance objectives.

The proposed outline includes eight major sections, starting with the scope and purpose of the standard and continuing through foundational concepts, categories of TEVV activities, and process recommendations. The draft emphasizes lifecycle approaches and iterative evaluation methods to ensure AI systems perform reliably and ethically over time.

The initiative stems from NIST’s broader efforts to accelerate AI standards development and increase stakeholder participation. NIST launched the Zero Drafts pilot in March 2025 following extensive engagement with industry, academia, and civil society. The TEVV draft is the first in a series, with a second outline on documentation of AI models and datasets expected soon.

“This outline represents a key milestone in building common technical language and guidance for evaluating the safety and reliability of AI,” said Elham Tabassi, Associate Director for Emerging Technologies at NIST. “Robust public input will help ensure the resulting standards are technically sound, globally relevant, and responsive to real-world challenges.”

Comments

Comments may be submitted via email to [email protected] and will be included in the public record. NIST plans to publish a full draft based on feedback and eventually submit the standard to a private sector-led standards body for formal development into a voluntary consensus standard.

Need Help?

If you’re wondering how these measures, or any other AI regulations and laws worldwide could impact you and your business, don’t hesitate to reach out to BABL AI. Their Audit Experts can address your concerns and questions while offering valuable insights.