Causal AI is within reach. It has the potential to trigger nothing less than a conceptual revolution in the law. This Essay explains why and takes a cautious look into the crystal ball. Causation is an elusive concept in many disciplines—not only the law, but also science and statistics. Even the most up-to-date artificial intelligence systems do not “understand” causation, as they remain limited to the analysis of text and images. It is a long-standing statistical axiom that it is impossible to infer causation from the correlation of variables in datasets. This thwarts the extraction of causal relations from observational data. But important advances in computer science will enable us to distinguish between mere correlation and factual causation. At the same time, artificially intelligent systems are beginning to learn how to “think causally.”
Tort Law
Everyone owes each other a duty of ordinary care—but what is “ordinary”? How does one act reasonably to meet this burden? This Comment analyzes the current reasonable person standard for disabled plaintiffs and the corresponding duty of “ordinary care” provided by defendants through a critical disability studies lens. The current system of tort law burdens disabled plaintiffs with accommodating themselves, rather than requiring defendants to include accessible care in meeting their duty of ordinary care. To make the distribution of accommodative labor more equitable, this Comment proposes three stackable changes: (1) courts should reinterpret defendants’ duty of ordinary care to include care of individuals with disabilities by eliminating the doctrine that tortfeasors owe accommodations to people with disabilities only if they are on notice of their disabilities; (2) courts could further shift the balance of accommodative labor by factoring the mental and physical cost of accommodating oneself into the reasonable care inquiry when the plaintiff is disabled; and (3) courts could eliminate comparative negligence for plaintiffs with disabilities to address the problematic “reasonable person with a disability” standard. This Comment also explores theoretical, doctrinal, and normative justifications while creating space for a more robust dialogue on how the law treats disability as “extra”—but not ordinary.
This paper examines the distinct features of artificial intelligence (AI) and reaches a broader conclusion as to the availability and applicability of first-order tort rules. It evaluates the accuracy of the argument that AI is similar in essence to other emerging technologies that have entered our lives since the First Industrial Revolution and, therefore, does not require special legal treatment. The paper will explore whether our current tort doctrines can serve us well even when addressing AI liability.
tandard economic analysis views strict liability as preferable to negligence because it is easier to administer and leads to better risk reduction: strict liability induces injurers not only to optimally invest in precaution but also to optimally adjust their activity levels. Standard analysis thus views the prevalence of negligence as unjustifiable on efficiency grounds. This Article challenges the conventional wisdom and clarifies an efficiency rationale for negligence by spotlighting the information-production function of tort law.
A few days before Christmas 1924, William Markowitz sold an air rifle to Richard Kevans. Markowitz should not have made that sale.
Richard Posner was certainly the most able judge in the history of tort law and in the development and deployment of law and economics.