# Colorado Is Writing AI Transparency Rules. The FTC Says Federal Law May Override Them. > Source: > Published: 2026-08-22 02:16:55+00:00 A quiet but profound structural tension is currently defining the landscape of American artificial intelligence regulation. In Colorado, the state Attorney General’s office is moving forward with the rulemaking process for the [Automated Decision-Making Technology (ADMT) Act](https://leg.colorado.gov/bills/sb24-205), with public comments open until October 26, 2026. Simultaneously, federal authorities are signaling that such state-level mandates may be fundamentally incompatible with federal oversight. This friction creates a complex environment for builders and operators. While Colorado proceeds as if its regulatory framework will be the standard for AI accountability, the Federal Trade Commission (FTC) has introduced a competing legal theory that suggests state laws attempting to dictate AI outputs may be preempted by federal authority. This is not merely a procedural disagreement; it is a foundational conflict over who holds the power to define the rules of the road for the next generation of autonomous agents. The ADMT Act, enacted in 2024, is designed to govern AI systems that make consequential decisions. Its scope is broad, covering automated processes in housing, employment, insurance, education, healthcare, and government services. For developers of AI agents, this means that any system capable of performing tasks like insurance underwriting, hiring recommendations, or healthcare triage must comply with rigorous transparency and risk-mitigation requirements. The state has pushed the enforcement date to January 1, 2027, to allow for the development of these implementing rules, but the underlying mandate remains a significant compliance hurdle. The federal counter-argument arrived on July 1, 2026, when the FTC published a [policy statement](https://www.federalregister.gov/documents/2026/07/01/2026-15555/policy-statement-on-the-application-of-consumer-protection-laws-to-ai-systems) in the Federal Register. The commission asserted that state laws requiring AI companies to suppress or alter accurate outputs for ideological or political objectives may be preempted by federal consumer protection law under Section 5. The core of this theory, as articulated by FTC Commissioner Ferguson, is the concern that state-level mandates could force companies to make their systems less accurate or less useful, thereby undermining the broader goals of consumer protection. This preemption theory directly targets the logic of the ADMT framework. If a state regulation requires an AI to alter its decision-making process in a way that impacts the accuracy of its output, the FTC argues that federal law should take precedence. It is important to note that this remains a proposed legal position rather than settled law. No court has yet ruled on the validity of this preemption theory, leaving companies in a state of regulatory limbo. The federal government’s interest in this conflict is not purely theoretical. In April 2026, the Department of Justice (DOJ) intervened in *xAI v. Colorado*, providing a clear signal that the federal executive branch is closely monitoring the preemption question. This intervention suggests that the conflict between state-level AI governance and federal authority is a priority for the current administration. For the crypto and AI professional, the stakes are high. We are witnessing a global transparency inflection point, marked by the recent enforceability of the [EU AI Act’s Article 50](/learn/what-is-the-eu-ai-act/) and the simultaneous implementation of California’s SB 942. As these regimes overlap, the question of who defines AI transparency obligations becomes paramount. If state laws are preempted, the regulatory burden may shift entirely to federal agencies, potentially creating a more uniform, albeit potentially more restrictive, environment. The current rulemaking in Colorado is a test of state sovereignty in the age of algorithmic decision-making. However, the FTC’s stance serves as a reminder that the federal government is unlikely to cede control over the national AI infrastructure. As the October 26 comment deadline approaches, the industry must navigate a path between complying with state-level mandates and preparing for a potential federal override. The outcome of this tension will likely determine the future of AI deployment in the United States for years to come.