How AI Is Shaping U.S. Regulatory Policies and Laws

Artificial intelligence is now playing a central role in shaping U.S. regulatory policies. As government agencies increasingly adopt AI tools, the technology is influencing decision-making processes and transforming how regulations are designed and enforced across the nation.

AI at the Heart of Federal Regulatory Drafting

A striking shift is underway within the U.S. Department of Transportation (DOT): regulatory texts once crafted by human experts are now being drafted using Google Gemini, a leading-edge generative artificial intelligence (AI). This move, championed by the administration of Donald Trump, aims to reshape the very mechanics of rulemaking at the federal level.

A Drive for Efficiency—At What Cost?

Proponents inside the DOT claim that harnessing AI could dramatically accelerate the creation and review of essential regulations. According to reported remarks from Gregory Zerzan, general counsel at the department, obtaining a draft regulation with Gemini should “not take more than 20 minutes.” In fact, internal presentations reportedly went as far as to call some current regulatory language a mere “word salad,” suggesting that automation could hardly do worse.

Several factors explain this decision:

• A desire to cut time-consuming bureaucratic processes;
• An acute shortage of staff, with over 4,000 employees—including about a hundred lawyers—departing since Trump’s second term began;
• A presidential mandate to modernize federal agencies without delay.

Skepticism and Warnings from Experts

However, not everyone in Washington is convinced. The stakes are especially high in transportation, where every word can have legal or life-and-death consequences—from aviation safety to hazardous material handling. Critics stress that while AI has aided with translation or data analysis in government work, actually authoring regulatory text is an unprecedented leap.

Skeptics abound. Mike Horton, former acting head of AI at the DOT, compared relying on Gemini to letting a high school intern draft legislation. Meanwhile, academic voices such as Bridget Dooling of Ohio State University urge caution: “It’s tempting to experiment with these tools,” she notes, “but skepticism is essential.”

The Broader Implications

Mr. Zerzan himself acknowledged that perfection is not the goal; rather, he suggests producing something “good enough” for immediate sector-wide impact. For observers, this attitude raises alarms about potential legal vulnerabilities or even real-world disasters stemming from hasty or flawed rulemaking.

Washington’s experiment signals an era where efficiency may clash with longstanding safeguards in public administration—a gamble whose outcomes remain highly uncertain as federal agencies test the boundaries of what AI can responsibly deliver.