M-21-06 (November 2020, still in force and cross-referenced by M-25-21) directs agencies, when considering regulatory or non-regulatory approaches to AI applications outside government, to avoid unnecessary barriers to innovation and to apply ten principles: public trust in AI, public participation, scientific integrity and information quality, risk assessment and management, benefits and costs, flexibility, fairness and non-discrimination, disclosure and transparency, safety and security, and interagency coordination; to consider non-regulatory approaches such as sector-specific guidance, pilot programmes and voluntary consensus standards; to follow the rulemaking guidance in its Appendix A; and to submit and maintain the agency plan on AI regulation described in Appendix B (statutory authorities, information collections, use case priorities, planned and existing regulatory actions).
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.