Donald Trump's announcement of an "AI Force" and forthcoming "AI Czar" arrives wrapped in the language of action but stripped of every detail that would make it actionable. No budget, no statutory authority, no mandate, no timeline beyond "in the near future." The post explicitly promises not to slow industry growth, then gestures vaguely at using "already existing Criminal and Civil Justice System" to catch bad actors. This is not a regulatory framework. It is the announcement of an announcement. The timing is not accidental. California Governor Gavin Newsom had just unveiled plans for a commission to regulate Silicon Valley AI companies, including a proposal for emergency kill switches on AI agents. Several Republican-led states are also pursuing regulation. Trump's move reads as federal pre-emption theater — asserting jurisdiction without exercising it, creating the impression of oversight that could later be used to argue state regulation is redundant. Meanwhile, the US-China AI competition provides the strategic justification for inaction. Anthropic CEO Dario Amodei's widely discussed essay frames the dilemma starkly: if the US restrains its AI capabilities and China does not, the power imbalance could become "militarily existential." This argument has become the load-bearing wall for the deregulatory position — any constraint on US companies is reframed as a gift to Beijing. The space race analogy is doing heavy rhetorical work, even as the actual policy environment looks nothing like NASA's centrally directed programs. China's approach to AI regulation is real but serves a fundamentally different purpose. Enforced through the Cyberspace Administration, Chinese rules ensure AI outputs align with state political doctrine — DeepSeek's regime-loyal answers on Taiwan and Tiananmen are features, not bugs. China's 2017 State Council development plan and its 2030 dominance timeline are industrial policy instruments, not safety frameworks. The US debate conflates these two entirely different regulatory logics. The upcoming Trump-Xi state dinner puts AI squarely on the bilateral agenda. US Treasury Secretary Scott Bessent has already proposed an information-exchange mechanism on serious AI risks, framing it as a transparency initiative between "the number one and the number two AI powers." But transparency on AI risks sits in a negotiating queue alongside trade restrictions on high-performance AI chips, rare earth elements, and extension of a basic trade agreement. Any AI concession will be traded against these harder economic interests. The structural pattern is clear: the US is building an AI governance apparatus that is deliberately hollow on the domestic side while pursuing bilateral frameworks with China that are aspirational at best. Industry captures the benefits of government-backed legitimacy and strategic urgency without bearing meaningful oversight costs. The public bears the risk of unregulated deployment while state-level attempts at consumer protection face federal pre-emption. If this pattern holds for two decades, the US ends up with an AI sector governed by voluntary industry commitments, occasional prosecutions after harm occurs, and a patchwork of state laws that federal courts systematically weaken. The comparison to financial services deregulation before 2008 is uncomfortable but structurally precise.
