The TAG Act (S. 5648) requires federal agencies to increase transparency and accountability when using artificial intelligence systems to make decisions affecting the public. The law defines key terms such as "automated systems," "critical decisions" (those affecting education, employment, benefits, housing, and healthcare access), and "plain language" communications, and applies specifically to AI systems that substantially influence agency decisions while excluding simple rule-based processes. Within 90 days of enactment, agencies must publish inventories of their AI systems and issue guidance on protecting privacy and civil rights, notifying the public when AI makes decisions affecting them, establishing appeal processes for adverse AI decisions, and tracking AI accuracy and fairness through performance metrics. The law also requires agencies to disclose when users are interacting with AI chatbots or similar systems and to provide users with complete transcripts of their AI interactions within three business days upon request. The TAG Act will automatically expire 10 years after its guidance takes effect unless Congress votes to reauthorize it.