For decades, government officials and members of the public have attempted to reschedule marijuana, which involves moving it from Schedule I of the Controlled Substances Act (CSA) to a less-tightly regulated category. Proponents gained momentum when marijuana reform became a focus of the 2024 U.S. presidential election. Republicans and Democrats pledged to decriminalize marijuana, legalize it, or end unnecessary arrests. Substantial confusion followed. People wondered whether rescheduling could legalize marijuana or decriminalize it, and some stakeholders claimed rescheduling would put the U.S. out of compliance with international drug treaties. In 2026, the Department of Justice ordered a narrow range of marijuana products to be recategorized as Schedule III. Although historic, the order did little to reduce confusion. Instead of harmonizing state and federal drug laws, it further fractured the U.S. drug policy landscape. Dozens of U.S. states have legalized or decriminalized marijuana for medical or nonmedical use, and policymakers are considering similar changes for other controlled substances. The DOJ order split state marijuana industries in half because, while state-licensed medical marijuana products were rescheduled, unlicensed products, as well as marijuana used for nonmedical purposes, remained in Schedule I. This Essay disentangles drug rescheduling, decriminalization, and legalization. Decriminalizing drugs eliminates or reduces criminal penalties for conduct associated with their personal use. Legalization creates regulatory frameworks governing drug production, distribution, and sale. The Essay explains how this distinction affects federal preemption, international treaty compliance, and public health. It applies the Supreme Court's anticommandeering framework from Murphy v. NCAA to show that decriminalization is constitutionally protected because Congress cannot compel states to criminalize conduct. Legalization, by contrast, often creates affirmative regulatory frameworks for conduct that federal law prohibits, raising conflicts with the CSA. The Food, Drug, and Cosmetic Act (FDCA) can also preempt state laws that legalize federally controlled substances. This preemptive potential persists even if drugs are descheduled, which removes them entirely from CSA control. Both the CSA and FDCA provide domestic regulatory infrastructure through which the United States complies with international drug control treaties. The CSA regulates controlled substance manufacturing, limits production, and penalizes drug trafficking. The FDCA provides import controls, manufacturing standards, and distribution oversight. Consequently, state laws that conflict with either statute may therefore implicate the federal government's authority over foreign affairs, weakening the presumption against preemption. Rescheduling marijuana or other Schedule I controlled drugs does not legalize them or decriminalize them because FDCA constraints remain. However, descheduling can effectively decriminalize controlled substances, and both the U.S. Constitution and international drug treaties allow it- making decriminalization legally defensible across multiple domains, while legalization faces legal risk that varies with the degree to which state regulations undermine federal and international objectives. The Essay concludes with recommendations to reduce the confusion and fractured policy landscape associated with federal marijuana rescheduling.
This Essay argues that communicative platforms' epistemic power is relational, fluctuating between weak, proxy, and strong depending on their infrastructural power and their relationships with states, experts, and publics. Platforms draw on and undermine other actors' authority: they borrow legitimacy from external institutions in response to political pressure and they validate or marginalize epistemic authority using platform design and algorithms. This relational framing of platforms' epistemic power calls for policy solutions beyond the platform. It highlights relationships that limit, support, or enable platform power. Restoring epistemic authority to institutions requires both constraining platforms' power and addressing the legitimate distrust of communities historically excluded from these institutions. It requires supporting and reforming institutional power and expertise and managing platforms' relationship with these institutions.
The proliferation of software-based tools to generate scientific evidence for criminal trials raises serious questions of how to validate the reliability of software-based evidence. Yet, a broad survey of case law indicates that courts often avoid serious scrutiny of the software tools themselves. We argue that this omission points to a gap in scientific evidence doctrine, which currently focuses judicial inquiry on validation of (1) basic scientific principles and (2) the forensic technician's application of those principles. Software presents a third, intermediate component: implementation of software code. Validating the software developer's work is a task that is distinct from either validating the underlying scientific principles or validating the forensic technician's use of the software-based tool. We argue that an adversarial process is essential to proving the reliability of software-based tools. As a starting point, courts could take more responsibility for scrutinizing the software implementation aspects when considering admissibility of software-based evidence. But we go further here and suggest that the task of validating evidentiary software is uniquely suited to a mandatory "red-teaming" adversarial certification process. Centralized evaluation of the software aspects offers potential to bring greater systematic improvements to the forensic evidence domain.
Law generates data. And across a wide range of topics, scholars have explored how. Recently, commentators have noted that, despite its data-production function, law often pays insufficient attention to choices about the information it produces, creating accidental spillovers. This Essay adds to these accounts by arguing that, even when law intentionally produces data, how data are presented can have significant unintended, distortionary effects. For example, employers are required to submit data to regulators on workplace safety but may characterize incidents in ways that avoid scrutiny. Financial firms may also use complex and dense language when reporting poor earnings, making it more difficult for investors to understand the firm's performance. In other words, even where law expressly requires the production of information, those that produce it can distort what they report. The result: Data presentation can undermine the goals of data production. This Essay explains these distortions in data presentation-what it calls data distortions-using the novel example of medical device adverse event reporting. It shows that manufacturers have economic incentives to distort mandatory adverse event reports in ways that make it more costly for regulators to detect safety concerns. While some firms may distort data to avoid regulatory detection, others may lack incentives to correct distortions that arise innocuously. However they occur, data distortions can undermine the purpose of producing the information in the first place. In the case of medical devices, distortions make safety signals difficult to detect, delaying or eliminating regulatory actions that protect patients. If regulators cannot detect safety problems at all or in a timely manner, the data collected cannot serve their legal purpose: to act as an early warning system for unsafe devices. To reduce data distortion and patient harm, this Essay evaluates various legal, regulatory, and market-based reforms. It concludes that some mixture of these reforms can reduce data distortions in medical device adverse event reporting.
