Under Executive Order 14319 and M-26-04, every solicitation or order for a large language model issued after 11 December 2025 carries contract requirements on compliance with the two Unbiased AI Principles (truth-seeking and ideological neutrality), as do solicitations for other AI models the agency decides the requirements should reach (considering the Appendix A.2 factors for agency-developed models and non-LLM generative AI); existing LLM contracts are modified to add the requirements as far as practicable and at the latest before any option extending performance is exercised. National security systems are outside the memorandum, open-source LLMs are exempt from the contractual requirements but due diligence on alignment is expected, and the memorandum sunsets two years after issue unless OMB extends it.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.