The federal government is urged to implement regulatory sandboxes to maintain the United States’ leadership in artificial intelligence (AI) development. As President Donald J. Trump considers an executive order that could challenge state laws exceeding their constitutional authority to regulate AI, focus shifts to proactive measures Congress can take to support evidence-based AI governance. Regulatory sandboxes, which allow for trial-and-error experimentation with new technologies, emerge as a promising strategy to facilitate this governance.
Regulatory sandboxes formalize an approach that has historically proven effective in managing novel technologies: learning through experimentation. As one observer noted, “one could well ask whether any technology, including the most benign, would ever have been established if it had first been forced to demonstrate that it would do no harm.” The introduction of new technology invariably carries risks, and society must strive to mitigate these while balancing the benefits. In the case of AI, despite a tendency among academics to focus on its downsides, the technology’s benefits warrant a trial-and-error approach through such regulatory frameworks.
Typically, regulatory sandboxes allow participants to deploy their products with minimal legal limitations. However, such participation is not a free pass for reckless behavior; it usually requires increased oversight, information-sharing obligations, and adherence to specific consumer protections. While some states have adopted the “try-first” approach to AI governance encouraged by Trump, others have instead imposed vague and subjective conditions that hinder effective experimentation.
This patchwork of state regulations poses significant challenges, as highlighted by Representative Ted Lieu (D-Calif.) during a recent congressional hearing on AI governance. He noted that differing standards across states could create compliance nightmares for labs, making it impractical for them to train models under multiple regulatory regimes. The failure to maintain a unified regulatory framework could put the U.S. at a disadvantage in both economic and national security terms, especially as adversaries continue to advance their own AI initiatives without waiting for the U.S. to catch up.
Without federal preemption, AI labs may struggle to fully utilize the regulatory sandboxes offered by a few forward-thinking states. Technology companies often gravitate toward the strictest jurisdiction’s laws, which stifles innovation in states that have embraced a more experimental regulatory approach, such as Utah and Texas. In contrast, states like California and New York have enacted regulations that reflect a fear of speculative risks, potentially hindering technological progress.
Congress could address this issue by establishing a legal safe harbor for AI labs participating in a federal regulatory sandbox or a state sandbox that meets certain criteria. Such a safe harbor would prevent individual states from obstructing the broader benefits of existing sandboxes, allowing for a more comprehensive understanding of AI’s advantages and risks. Another potential route could involve a temporary moratorium on certain state laws, making participation in existing sandboxes more feasible.
Both proposed actions would not eliminate the prospect of future regulations on AI but rather support informed discourse on the subject. This approach aims to ensure that biased academic and media narratives do not cloud legislative decisions. Attempts to overly restrict AI or impose numerous procedural checks will not facilitate a full understanding of its associated risks and benefits. This perspective aligns with the evidence-based governance model advocated by AI experts convened by California Governor Gavin Newsom.
Time is of the essence. The longer state-level AI regulations remain in effect, the more likely labs will adapt their operations to comply with them, potentially stunting the pace and direction of AI innovation for years to come. It is crucial for Congress to act swiftly and decisively to secure the U.S. position at the forefront of AI advancement.
See also
China Proposes World AI Cooperation Organisation to Set Global Safety Standards
FTC Warns Businesses: Key Legal Risks of AI Chatbots Require Immediate Safeguards
Federal AI Regulation Moratorium Reemerges Amid Bipartisan Opposition from States
Hochul Revises AI Safety Bill, Aligns with Big Tech Interests Amid Lobbying Pressure
Trump Signs Executive Order Limiting State AI Regulations to Boost U.S. Innovation
