Congress remains deadlocked over a national artificial intelligence (AI) framework, but a patchwork of aggressive state laws is slowly transforming the U.S. regulatory landscape.
For now, landmark legislation in California, Colorado, and Illinois is setting a high-stakes precedent for the tech industry.
With over 1,500 AI-related bills introduced this year alone, the legal environment has split into three distinct fronts: safety protocols for frontier models, the elimination of algorithmic bias in hiring, and a mandatory “right to know” for consumers interacting with synthetic media.
California remains the epicenter of the crackdown.
Following the high-profile veto of previous kill switch legislation, the state moved to Senate Bill 53, which took effect on Jan. 1, 2026. This law targets the industry’s most powerful systems—those trained with massive computational power. Developers are now legally required to publish detailed risk frameworks and report any “critical safety incidents” to state regulators.
Simultaneously, Assembly Bill 2013 has pulled back the curtain on the black box of AI, forcing companies to disclose the datasets used to train their models, including the use of copyrighted material.
With no federal legislation in sight, state laws are providing a template for citizens and companies though ultimately navigating AI requires a partnership between government and education, according to Darrell Steinberg, the former mayor of Sacramento, Calif., and leader of the majority party (Democratic) in the California State Senate from 2008 to 2014.
“We’re at an existential point with AI. It may change how we define work or the (shorter) work week,” he said in an interview. “There is a cultural separation between education and the workforce. The educational system needs to move to more practical experience through internships and apprenticeship programs.”
While California monitors AI models, Colorado and Illinois are focusing on the outcomes. On June 30, 2026, the Colorado AI Act (SB 24-205) will reach full implementation, representing the nation’s first comprehensive assault on high-risk AI. The law targets systems used for consequential life decisions—such as housing, healthcare, and lending—requiring annual impact assessments to detect discrimination. Crucially, it grants citizens the “right to appeal” an AI’s decision to a human representative.
In a similar vein, Illinois recently enacted HB 3773. Effective since the start of the year, the law treats discriminatory AI in the workplace as a civil rights violation. It specifically bans the use of “proxy data,” such as zip codes, which AI systems often use to bypass traditional anti-discrimination protections.
Perhaps the most visible shift for the public involves the fight against deepfakes. Under California’s SB 942, which becomes enforceable this August, platforms must provide tools to detect AI-generated content and embed latent metadata watermarks in all outputs. This push for transparency is mirrored in states like Montana and Nevada, where new 2026 statutes provide residents with “publicity protections,” allowing them to sue for damages if their likeness is used in synthetic media without consent.
This legislative surge has not gone unnoticed by the executive branch. A late 2025 Executive Order has already tasked the Department of Commerce with identifying state laws that may “conflict with federal policy” or infringe on First Amendment rights.
As state attorneys general gear up for a defense of their local mandates, this year is shaping up to be a defining courtroom battleground for the future of American innovation and civil liberty.
The industry, meanwhile, is intent on self-policing its AI use. For example, OpenAI recently proposed a technological New Deal of sorts.
Its 13-page policy paper, “Industrial Policy for the Intelligence Age: Ideas to Keep People First,” proposes the creation of a National Public Wealth Fund. To fund this and protect a tax base potentially hollowed out by automation, OpenAI proposes a “robot tax,” levies on companies that significantly reduce costs by replacing human labor with automated systems; moving from payroll taxes toward higher capital gains and corporate income taxes; and using fund returns to provide “citizen dividends,” ensuring the AI boom doesn’t just enrich a handful of tech giants.
