According to CoinMeta, there is a significant gap between the voluntary commitments of the White House regarding the AI protocol and the mandatory enforcement measures in Connecticut. On September 28th, NVIDIA announced the launch of a proxy security platform ( OASP), indicating that the industry prefers self-governance over government regulation. Subsequently, on September 29th, the White House signed a joint commitment on responsible innovation, involving companies such as Google, Anthropic, Meta, OpenAI, XAI, and NVIDIA. Although this agreement lists four layers of control measures, they are entirely voluntary and non-binding, lacking an enforcement mechanism and penalty provisions. On October 1st, Connecticut's AI Responsibility Act ( CAIA ) came into effect; it is mandatory and requires developers to share information and make specific disclosures, with violators facing fines of up to $10,000 per instance. Unlike the federal agreement, the Connecticut framework is legally binding, highlighting a structural disparity in governance between the federal and state levels regarding AI.