Federal agencies must complete AI security audits by July 15 under a new directive from the White House.
The audits focus on model integrity, data leakage, and adversarial robustness for systems handling sensitive data.
Contractors supporting federal AI deployments will face corresponding requirements later this year.
