The **Department of Defense**'s ongoing struggle with **Anthropic** over AI usage for mass surveillance highlights a critical debate. Now, the **GSA** is attempting to rewrite government procurement rules, potentially exacerbating these concerns. ### Procurement as Policy Using procurement to achieve policy goals is a common practice. Governments allocate resources to projects aligning with public interest, such as open-source software or right to repair, while avoiding those with inadequate security, as seen with shady contractors and their data breaches. ### GSA's New Guidelines The **GSA**'s proposed rules, intended to promote "ideologically neutral" American AI innovation, go far beyond. Comments filed by the **Center for Democracy and Technology**, the **Protect Democracy Project**, and the **Electronic Privacy Information Center** reveal provisions that could make AI tools less safe and less useful. You can read the full comments [here](https://www.eff.org/document/gsar-552239-7001-basic-safeguarding-artificial-intelligence-systems). ### "All Lawful Purposes" A key concern is the requirement that contractors license their AI systems to the government for "all lawful purposes." Given the government's history of expansive legal interpretations and surveillance practices, this raises serious privacy risks. The potential for misuse of personal data is significant, even for routine bureaucratic data. ### Disabling Safety Guardrails The draft rules also mandate that "AI System(s) must not refuse to produce data outputs or conduct analyses based on the Contractor’s or Service Provider’s discretionary policies." This means companies may be forced to disable safety guardrails if the government requests data that those guardrails would otherwise block. Limiting necessary safeguards is a misguided approach, given public concerns about AI safety. ### Other Issues Beyond these major issues, the draft rules contain other problematic elements, including technologically incoherent "anti-Woke" requirements. The overarching concern is that the proposal fails to promote privacy, safety, and responsible technological innovation. Critics are urging the **GSA** to reconsider and revise the rules to better serve the public interest.