No Adversarial AI Act

The "No Adversarial AI Act" aims to safeguard the United States Government from artificial intelligence developed by foreign adversaries. It directs the Federal Acquisition Security Council to establish and regularly update a public list identifying such AI, with the Office of Management and Budget responsible for its publication within 180 days of enactment. This list will be updated every 180 days and includes a process for AI to be removed if certified not to be foreign adversary AI. The legislation mandates that U.S. executive agencies review and consider the exclusion or removal of any artificial intelligence found on this list from their systems within 90 days. Agencies are generally prohibited from acquiring or using these identified foreign adversary AI products, leveraging existing authorities to mitigate associated risks. This ensures a proactive approach to protecting government operations from potentially compromised technology. However, the bill provides for specific exceptions to this prohibition, allowing an agency head to approve the acquisition or use of such AI under certain conditions. These conditions include necessity for scientifically valid research , evaluation, training, testing, or analysis , counterterrorism or counterintelligence activities , or to avoid jeopardizing mission-critical functions . Such exceptions require proper written notice to Congress and the Office of Management and Budget.