The "Validation and Evaluation for Trustworthy Artificial Intelligence Act" sponsored by Sen. Hickenlooper aims to develop new guidance for AI evaluations to ensure accurate testing and safe deployment of automated systems.
The legislation would involve NIST in creating consolidated approaches for monitoring risk levels, evaluating system efficacy, and setting standards for internal and external auditors to ensure AI systems work safely and as intended.
NIST, in collaboration with the NSF and DOE, would focus on safeguarding consumer privacy, mitigating system harms, ensuring dataset quality, disclosing updates, and maintaining governance controls to develop testing guidance for AI systems.
Another provision in the bill involves the Department of Commerce establishing an "Artificial Intelligence Assurance Qualifications Advisory Committee" to help execute the guidance and set standards for accrediting independent AI auditing organizations.
Collection
[
|
...
]