This Essay explores the future of legal prediction markets. Part I explains how markets work and what makes them hard to beat. Part II then turns to the largest legal prediction market to date: the outcome of Learning Resources. Finally, Part III considers whether markets are well-suited to forecasting legal outcomes, both in principle and in practice.
The central concern of structural constitutional law is the organization of governmental power, but power comes in many forms. This Article develops an original account of data’s structural law—the processes, institutional arrangements, transparency rules, and control mechanisms that, we argue, create distinctive structural dynamics for data’s acquisition and appropriation to public projects. Doing so requires us to reconsider how law treats the category of power to which data belongs. Data is an instrument of power. The Constitution facilitates popular control over material forms of power through distinctive strategies, ranging from defaults to accounting mechanisms. Assessing data’s structural ecosystem against that backdrop allows us to both map the structural law of data and provide an initial diagnosis of its deficits. Drawing on our respective fields—law and computer science—we conclude by suggesting legal and technical pathways to asserting greater procedural, institutional, and popular control over the government’s data.
Critics of generative AI often describe it as a “plagiarism machine.” They may be right, though not in the sense they mean. With rare exceptions, generative AI doesn’t just copy someone else’s creative expression, producing outputs that infringe copyright. But it does get its ideas from somewhere. And it’s quite bad at identifying the source of those ideas. That means that students (and professors, and lawyers, and journalists) who use AI to produce their work generally aren’t engaged in copyright infringement. But they are often passing someone else’s work off as their own, whether or not they know it. While plagiarism is a problem in academic work generally, AI makes it much worse because authors who use AI may be unknowingly taking the ideas and words of someone else.
Disclosing that the authors used AI isn’t a sufficient solution to the problem because the people whose ideas are being used don’t get credit for those ideas. Whether or not a declaration that “AI came up with my ideas” is plagiarism, failing to make a good-faith effort to find the underlying sources is a bad academic practice.
We argue that AI plagiarism isn’t—and shouldn’t be—illegal. But it is still a problem in many contexts, particularly academic work, where proper credit is an essential part of the ecosystem. We suggest best practices to align academic and other writing with good scholarly norms in the AI environment.
Beware dark patterns. The name should be a warning, perhaps alluding to the dark web, the “Dark Lord” Sauron, or another archetypically villainous and dangerous entity. Rightfully included in this nefarious bunch, dark patterns are software interfaces that manipulate users into doing things they would not normally do. Because of these First Amendment complications, the constitutionality of dark pattern restrictions is an unsettled question. To begin constructing an answer, we must look at how dark patterns are regulated today, how companies have begun to challenge the constitutionality of such regulations, and where dark patterns fall in the grand scheme of free speech. Taken together, these steps inform an approach to regulation going forward.
Holders of patents covering technology standards, known as standard-essential patents (SEP), control the rights to an invention with no commercially-viable alternative or that cannot be designed around while still complying with a standard. This gives SEP holders significant leverage in licensing negotiations. Standards development organizations (SDOs) play an important role in curbing opportunistic behavior by patent holders. SDOs require SEP holders to license their patents on fair, reasonable, and non-discriminatory (FRAND) terms. However, courts have mischaracterized FRAND commitments, concluding that these disputes carry a Seventh Amendment guarantee to a jury trial. This mischaracterization undermines the fair resolution of FRAND disputes, and a different approach is necessary. In this Comment, Marta Krason proposes an alternative analytical framework that more accurately characterizes FRAND disputes by drawing on principles from contract and property law, concluding that the constitutionally proper adjudicator is a judge, not a jury.
The internet plays a crucial role in modern life; however, equal access to it is not guaranteed. Drawing on existing tribal spectrum sovereignty arguments, Morgan Schaack writes that the control exercised by the FCC’s licensing of the electromagnetic spectrum and language common in many tribal treaties create a tribal access right to spectrum under the trust responsibility. Framing this access to spectrum as a trust-protected resource, the Comment situates allowing tiered internet service in the absence of net neutrality as a violation of the government's obligations under the trust responsibility.
Recently, many states have reacted to the growing data economy by passing data privacy statutes. These follow the “interaction model”: they allow consumers to exercise privacy rights against firms by directly interacting with them. But data brokers, firms that buy and sell data for consumers whom they do not directly interact with, are key players in the data economy. How is a consumer meant to exercise their rights against a broker with an “interaction gap” between them?
A handful of states have tried to soften the interaction gap by enacting data-broker-specific legislation under the “transparency model.” These laws, among other things, require brokers to publicly disclose themselves in state registries. The theory is that consumers would exercise their rights against brokers if they knew of the brokers’ existence. California recently went further with the Delete Act, providing consumers data-broker-specific privacy rights.
Assembling brokers’ reported privacy request metrics, this Comment performs an empirical analysis of the transparency model’s efficacy. These findings demonstrate that the transparency model does not effectively facilitate consumers in following through on their expected privacy preferences or meaningfully impacting brokers. Therefore, regulators should follow in the footsteps of the Delete Act and move beyond the transparency model.
This Essay argues for the development of more robust—and balanced—law that focuses not only on the risks, but also the potential, that AI brings. In turn, it argues that there is a need to develop a framework for laws and policies that incentivize and, at times, mandate transitions to AI-based automation. Automation rights—the right to demand and the duty to deploy AI-based technology when it outperforms human-based action—should become part of the legal landscape. A rational analysis of the costs and benefits of AI deployment would suggest that certain high-stakes circumstances compel automation because of the high costs and risks of not adopting the best available technologies. Inevitably, the rapid advancements in machine learning will mean that law soon must embrace AI; accelerate deployment; and, under certain circumstances, prohibit human intervention as a matter of fairness, welfare, and justice.
For data, the whole is greater than the sum of its parts. There may be millions of people with the same birthday. But how many also have a dog, a red car, and two kids? The more data is aggregated, the more identifying it becomes. Accordingly, the law has developed safe harbors for firms that take steps to prevent aggregation of the data they sell. A firm might, for instance, anonymize data by removing identifying information. But as computer scientists have shown, clever de-anonymization techniques enable motivated actors to unmask identities even if the data is anonymized. Data brokers collect, process, and sell data. Courts have traditionally calculated data brokering harms without considering the larger data ecosystem. This Comment suggests a broader conception is needed because the harm caused by one broker’s conduct depends on how other brokers behave. De-anonymization techniques, for instance, often cross-reference datasets to make guesses about missing data. A motivated actor can also buy datasets from multiple brokers to combine them. This Comment then offers a framework for courts to consider these “network harms” in the Federal Trade Commission’s (FTC) recent lawsuits against data brokers under its Section 5 authority to prevent unfair acts and practices.
When the past is thought to predict the future, it is unsurprising that machine learning, with access to large data sets, wins prediction contests when competing against an individual, including a judge. Just as computers predict next week’s weather better than any human working alone, at least one study shows that machine learning can make better decisions than can judges when deciding whether or not to grant bail.
