Senator Ted Cruz Proposes AI ‘Sandbox’ to Ease Federal Regulations
U.S. Senator Ted Cruz on Wednesday introduced a bill that would create a regulatory “AI sandbox” allowing artificial intelligence companies to apply for temporary exemptions from certain federal rules while developing new technologies.
Cruz, who chairs the Senate Commerce Committee, described the proposal as a way to help U.S. firms stay competitive with China by lowering regulatory barriers. “A regulatory sandbox is not a free pass. People creating or using AI still have to follow the same laws as everyone else,” Cruz said during a subcommittee hearing.
Key Details
-
The bill would let federal agencies grant two-year exemptions to companies that apply, provided they outline safety and financial risks and how they would mitigate them.
-
The Office of Science and Technology Policy (OSTP) would be given authority to override agency denials of waivers.
-
The sandbox would apply only at the federal level — Cruz’s proposal does not preempt state-level AI regulations, despite pressure from the tech industry.
Industry Push and Opposition
Major AI developers including OpenAI, Google, and Meta have urged the Trump administration to reduce regulatory barriers. The White House OSTP has also begun seeking public input on which regulations hinder AI growth.
Consumer advocacy group Public Citizen sharply criticized Cruz’s bill, arguing it “treats Americans as test subjects” and warning against OSTP’s ability to override regulators. “The sob stories of AI companies being ‘held back’ by regulation are simply not true,” said J.B. Branch, the group’s Big Tech accountability advocate, pointing to record-high valuations of AI firms.
State-Level Rules
While Cruz’s bill avoids limiting state laws, AI regulation is already expanding at the state level:
-
California bans unauthorized political deepfakes and requires patient disclosure when AI is used in healthcare.
-
Colorado passed a law to curb AI discrimination in hiring, housing, banking, and other areas — its enforcement was pushed to mid-2026 after lobbying by the tech sector.
-
Several states have criminalized AI-generated explicit imagery without consent.
OSTP director Michael Kratsios told the committee that such state measures risk stifling innovation, suggesting Congress revisit preemption in the future.
The proposal is likely to fuel debate between those who see regulation as a barrier to U.S. innovation and those who warn of the risks of treating AI experimentation as a public trial.



