This commentary was originally published by Notice & Comment.
Last summer, the Department of Housing and Urban Development (HUD) received a PowerPoint presentation introducing an Artificial Intelligence (AI) tool: SweetREX, named after its creator, a third-year undergraduate in economics. Consistent with the Trump administration’s stated goal of eliminating 50 percent of all federal rules by the first anniversary of President Trump’s inauguration, Elon Musk’s so-called Department of Government Efficiency (DOGE) had used to SweetREX to review more than a thousand of HUD’s regulatory sections and reportedly used it to write 100% of deregulations at the Consumer Financial Protection Bureau (CFPB). These were not simply idle recommendations for human decision makers to consider; rather, the AI recommendations were treated as presumptively correct, with staffers required to justify in writing any disagreement with what the model suggested.
The growing use of AI for such complex governance functions as regulatory decision-making is attracting attention from policy experts and policymakers alike. The example of HUD’s and the CFPB’s use of SweetREX, in particular, illustrates how such situations risk putting two existing bodies of public law—those governing the rulemaking process and procurement, respectively—on a collision course: in particular, rulemaking’s emphasis on transparency and procurement’s emphasis on quick commercial transactions that often do not include robust transparency audits.
The impetus for this potential clash is that the architecture of many of the AI systems that the federal government employs are developed by private contractors. For instance, rather than being developed entirely in-house, SweetREX is powered primarily by Google Gemini’s model. This means that the federal government is unlikely to have a full understanding of how the model was trained.
This gap in understanding is problematic in the context of agency rulemaking decisions because the law that governs these actions—the Administrative Procedure Act (APA)—requires the responsible agencies to offer an accounting of the deliberative process that undergirds their final decisions. Under that law’s “arbitrary-and-capricious” standard, which is used to evaluate the policy rationale for a regulation, the Supreme Court has held that agencies may not rely on factors Congress did not intend them to consider (i.e., “improper factors”). Of course, policing this standard is not easy even when agency decisionmakers are human. Hesitant to probe the minds of agency decisionmakers, courts have turned to the functional solution of evaluating the agency’s decision-making process.
However, AI complicates this analysis due to the “black box” problem. The inability to truly understand how an AI system reaches its decision, including what factors it weighs and how it weighs each factor. If the model relied on an improper factor, the agency that deployed it has effectively done so as well, with no way of knowing it has.
Showing 2,976 results
Arvind Salem, James Goodwin | June 8, 2026
The growing use of AI for such complex governance functions as regulatory decision-making is attracting attention. The practice risks putting two existing bodies of public law—those governing the rulemaking process and procurement, respectively—on a collision course: in particular, rulemaking’s emphasis on transparency and procurement’s emphasis on quick commercial transactions that often do not include robust transparency audits. The impetus for this potential clash is that the architecture of many of the AI systems that the federal government employs are developed by private contractors.
Daniel Farber | June 1, 2026
A proposed new rule would replace merit-based funding of science with a heavily politicized process. It would also deprive scientists of the ability to rely on finding for long-term projects. This proposal has no real legal basis. It is likely to undermine American science, giving a boost to China's efforts to take over global scientific leadership.
Alejandro Camacho, Robert L. Glicksman | May 21, 2026
This excerpt is from a commentary originally published in The Conversation. In recent years, at least two dozen local and state governments have sued petroleum companies to recover the billions in costs they have incurred responding to and rebuilding after flooding, storms and wildfires – all of which have been worsened by changes to the […]
Daniel Farber | May 15, 2026
EPA has said it would no longer try to quantify the harms done by the two most serious, widespread air pollutants. Given that these are the most fully understood of all environmental impacts, it’s not clear what future regulations, if any, might still be subject to cost-benefit analysis. This didn’t come out of the blue. Rather, it is the culmination of a series of steps that began when President Donald Trump first took office in 2017.
Federico Holm | May 12, 2026
It is tempting to think that the threat of the current Congress abusing the Congressional Review Act (CRA) is over, now that the deadline to revisit rules implemented during the previous Congress’s session—provided by the CRA’s unique “lookback provision”—has formally passed. But that would be a mistake, as conservative lawmakers have found novel ways to target agency actions from previous administrations.
Steph Tai | May 7, 2026
The question of which court should hear a case isn’t always as easy as it might seem — and the answer can sometimes make a difference in the potential outcome. For instance, in 2013, the government of Plaquemines Parish, Louisiana, decided to sue several oil companies for violating a 1978 state law that required a state permit for oil production along the Louisiana coast.
Alejandro Camacho | May 5, 2026
The Trump administration recently repealed the U.S. Environmental Protection Agency’s 2009 endangerment finding—the scientific and legal determination that greenhouse gases threaten public health and welfare that has anchored federal climate regulation for nearly two decades. Environmental Protection Agency Administrator Lee Zeldin called the finding “the Holy Grail of federal regulatory overreach.” Within weeks, a coalition of more than 20 states filed suit in the U.S. Court of Appeals for the D.C. Circuit to reverse the repeal. The legal battle that follows will help define American environmental policy for a generation. Brigham Daniels and I did not plan the timing of our new book, Lessons for a Warming Planet: A Vital History of US Environmental Law, to coincide with this particular legal conflict. But we could not have chosen a more clarifying moment for its release. The endangerment finding repeal is not an aberration—it is a recognizable recurrence in a history that stretches back centuries. Law has always been the primary engine of both environmental exploitation and protection in the United States.
Bryan Dunning | May 4, 2026
On April 23, the Baltimore Sun published an op-ed lambasting efforts by localities and states to hold the fossil fuel industry responsible for decades of misinformation about the dangers of their products. These hazards to the public’s health, welfare, and safety are now coming home to roost in the form of extreme weather, increased flooding […]
James Goodwin | May 4, 2026
On April 24, The Washington Postpublished an op-ed that sought to blame an unusual source for the high energy and gas prices Americans are now facing: justice. Specifically, it claims that these price increases are the result of state and local governments trying to hold Big Oil accountable for the climate-related harms their constituents are continuing to suffer. I submitted a letter to the editor debunking the op-ed’s argument, noting that the real cause of higher gas prices is the illegal war in Iran.