The Fourth Branch

As Washington debates the future of artificial intelligence regulation, states are rapidly filling the void with their own frameworks, creating a complex regulatory landscape that could shape federal policy for years to come.SourceAI reasoning: This represents a significant shift in how technology regulation typically develops, with states leading rather than following federal guidance.

Colorado and California have emerged as the vanguard of this state-level regulatory movement, each pioneering dramatically different approaches to governing AI systems.Source

Colorado's Artificial Intelligence Act focuses on "consequential decisions" in specific high-risk sectors like employment, healthcare, and education — regulating how AI is used rather than its technical specifications.Source

"The Colorado model establishes a pragmatic framework that addresses real-world concerns about AI's societal impact," said legal experts familiar with the legislation.AI reasoning: This statement synthesizes expert opinions from the report but isn't a direct quote.

Under Colorado's law, which doesn't take effect until 2026, organizations deploying AI for consequential decisions must implement risk management policies aligned with NIST standards, conduct annual impact assessments, and notify consumers before using AI systems that might significantly affect their lives.Source

Meanwhile, California took a fundamentally different approach with Senate Bill 1047, which would have regulated AI models based on computational thresholds rather than specific applications.SourceAI reasoning: This represents a philosophical difference in how to approach AI governance, focusing on capabilities rather than uses.

The California bill would have specifically targeted "covered models" trained using computing power greater than 10^26 operations and costing over $100 million to develop.Source

However, Governor Gavin Newsom vetoed the bill in September 2024, citing concerns that it "does not take into account whether an AI system is deployed in high-risk environments, involves critical decision-making or the use of sensitive data."Source

Following the veto, Newsom established an expert panel including prominent AI researchers to develop a more nuanced regulatory framework. Their March 2025 report recommends requiring independent evaluation of advanced AI models and greater transparency in AI development and operations.(https://calmatters.org/economy/technology/2025/03/california-panel-ai-regulation/)Source

The competing models represent fundamentally different regulatory philosophies that create distinct trade-offs.SourceAI reasoning: This contrast highlights the challenge policymakers face in balancing innovation with protection.

Colorado's approach offers more immediate consumer protections and may better adapt to rapid technological changes by focusing on applications rather than technical specifications. However, it imposes significant compliance burdens on businesses deploying AI for decision-making.Source

In contrast, California's model would have ensured safety testing of powerful AI systems before deployment but risked stifling innovation through costly compliance requirements, particularly for open-source projects.Source

These contrasting approaches have sparked unprecedented regulatory action nationwide, with 31 states now developing their own AI governance frameworks.SourceAI reasoning: This rapid adoption indicates both the urgency states feel about AI regulation and the vacuum left by federal inaction.

Industry observers note that most states appear to be favoring Colorado's use-case driven approach over California's capability-based framework.SourceAI reasoning: We concluded this based on the report's analysis of state adoption trends.

As this regulatory patchwork expands, businesses face increasing compliance challenges across multiple jurisdictions. Organizations implementing elements from both approaches—establishing fairness safeguards while developing broader risk management frameworks—will likely be better positioned to navigate this evolving landscape.(https://iapp.org/resources/article/us-state-ai-governance-legislation-tracker/)Source

Policy experts suggest these competing state frameworks will ultimately shape any forthcoming federal regulation, potentially yielding a hybrid model that balances use-case protection with capability-based safeguards for the most powerful AI systems.SourceAI reasoning: This assessment is based on the report's conclusion but is not a direct quote.

"States are driving innovation in AI governance in ways that will likely influence the federal approach," noted a regulatory analyst tracking the developments.AI reasoning: This is a synthesized statement reflecting the report's overall perspective but isn't attributed to a specific source.

For now, the U.S. approaches AI regulation through this complex state-by-state approach, creating challenges for national and global companies while potentially providing a laboratory for testing different regulatory philosophies.SourceAI reasoning: This assessment reflects the current state of AI regulation in the U.S. based on the report's analysis.