VET Artificial Intelligence Act

The VET (Validation and Evaluation for Trustworthy) Artificial Intelligence Act would require the Director of the National Institute of Standards and Technology (NIST) to create voluntary guidelines and specifications for both internal and external assurances of AI systems. These guidelines are designed to help developers, deployers, and independent evaluators test, validate, and verify AI systems to build trust, improve governance, and support accountability. The act emphasizes consensus-based, risk-aware approaches and alignment with NIST’s AI Risk Management Framework, while avoiding mandates to use particular technologies or products. In addition, the bill creates an advisory committee to study qualification and accreditation frameworks for assurance providers and mandates a study and report on the market and capabilities of entities that conduct AI assurances. Overall, the bill seeks to establish a voluntary, standards-aligned assurance ecosystem for AI, with processes for development, public input, and ongoing updates, plus governance structures to explore how assurances could be licensed, accredited, or otherwise validated in the future.