The General Services Administration (GSA) is introducing new contract guidelines for artificial intelligence (AI) vendors serving the federal government. These guidelines mandate that AI providers must grant agencies broad operational rights to use their models for any lawful purpose. The draft policy proposes that vendors provide an “irrevocable, royalty-free, non-exclusive license” for the government to utilize AI systems throughout the contract duration. It also restricts vendors from limiting how their AI models respond to government inquiries. This initiative follows a conflict involving the Department of Defense and Anthropic over technology usage policies. Additionally, the proposal emphasizes that AI outputs must align with historical accuracy and neutrality, rejecting ideological biases. Continuous monitoring for issues like bias and safety is also a key component, allowing the government to assess AI systems independently. However, civil liberties advocates argue that these demands could undermine essential safety measures, potentially deterring responsible vendors from engaging with the government.
Source link
Share
Read more