U.S. AI rulemaking splinters
AI governance in the U.S. is fragmenting: Elon Musk’s xAI sued Colorado to block a first‑in‑the‑nation state AI law set to take effect in June, arguing it would create liability and free‑speech issues. (nationaltoday.com) At the federal level, the Commodity Futures Trading Commission named an Innovation Task Force to shape AI and prediction‑market rules, showing agencies are carving out sectoral roles. (pymnts.com) Courts are also active: an appeals court recently denied Anthropic’s challenge in a dispute with the Trump administration, underscoring legal contests as a channel for setting AI limits. (rswebsols.com)
The United States is setting artificial intelligence policy in pieces, not through one national rulebook. (cftc.gov) On April 9, xAI sued Colorado in federal court to block Senate Bill 24-205, a 2024 state law scheduled to take effect on February 1, 2026, with some notice and disclosure duties beginning earlier. The company said the law would expose developers to liability for “algorithmic discrimination” and violate the First Amendment, due process, equal protection, and the Constitution’s limits on state interference with interstate commerce. (ppc.land) Colorado’s law applies to “high-risk” systems used in “consequential decisions” such as hiring, housing, lending, education, health care, insurance, and legal services. It requires developers and deployers to use reasonable care to avoid algorithmic discrimination and to provide notices, risk management, and impact assessments in specified cases. (bhfs.com) At the federal level, the Commodity Futures Trading Commission moved in the opposite direction from broad consumer protection and toward sector-specific oversight. On March 24, Chairman Michael S. Selig announced an Innovation Task Force to help build rules for crypto assets, artificial intelligence and autonomous systems, and prediction markets and event contracts. (cftc.gov) That means one agency is treating artificial intelligence as part of market plumbing rather than a general civil-rights issue. The task force is being led by Michael J. Passalacqua, a senior adviser to Selig, and the commission said it will work with the Innovation Advisory Committee on the framework. (cftc.gov; pymnts.com) Courts are shaping the map too. On April 8, a federal appeals court in Washington refused Anthropic’s request to temporarily block a Trump administration action that labeled the company a supply-chain risk, even as a separate San Francisco case had already produced narrower relief for Anthropic. (pbs.org; cnbc.com) Anthropic said in those cases that the administration was retaliating against the company over its efforts to limit how its models could be used. The administration said Anthropic was trying to dictate United States military policy, and the appeals panel said the balance of harms favored the government at that stage. (pbs.org; cnbc.com) Congress has spent more than a year discussing federal artificial intelligence guardrails without producing a single comprehensive statute that would preempt state laws like Colorado’s. In that gap, states, regulators, and judges are each writing narrower rules for their own lanes. (bhfs.com; cftc.gov) The next test is not theoretical. Colorado’s law is headed for enforcement fights, federal agencies are assigning themselves separate artificial intelligence portfolios, and companies are increasingly using lawsuits to decide which limits stick. (ppc.land; cftc.gov; pbs.org)
