Top Student Papers

The growing presence of artificial intelligence (AI) tools in society is driving a policy movement for the responsible development of AI. A principle that is often stated as necessary is explainability, however, the exact meaning of what constitutes explainable AI is left vague. Explainability is generally a concept aimed at elucidating the nature of machine learning (ML) models. Explainable ML has different meanings depending on the context in which it is required. In the healthcare context, some academics are calling for a high degree of ML explainability, i.e., at the model level, to promote patient safety and foster trust in AI. However, requiring explainable ML in healthcare to realize the goals of patient safety and ML accountability is an overstated solution to these goals that fails to consider the cost of explainability to innovation in the field of ML healthcare tools. Specifically, requiring model level explainability can limit the scope of intellectual property protections available to developers, which negatively influences developers’ incentive to develop ML healthcare tools.

“Explainability” as a regulatory requirement or policy objective should be replaced with specific objectives rather than left as a vague requirement open to interpretation. This paper explores the consequences of explainability to innovation by dissecting the concept of “explainability” into more specific definitions that reflect the true objectives of requiring explainability in the healthcare context. This approach reveals that while a balance between requiring explainable ML and developers’ intellectual property interests is possible, the reasons for requiring explainability are rebutted by evidence that explainable ML will negatively impact patient safety and provider adoption more than it will promote these objectives.

Social media misinformation is the biggest public health threat now and into the foreseeable future. This issue requires government action. In this paper, I will argue that, as part of a holistic plan to address misinformation on social media, the federal government has the power to enact legislation that requires algorithms to increase the reach of certain posts while limiting the reach of other health-related posts during a health emergency. In other words, the federal government can and should implement legislation that changes social media algorithms to push reliable non-government health information to the top of feeds before health misinformation.

This paper is divided into three parts to make that argument. In the first part, I consider how legislation should be designed which includes looking at how other countries have addressed health misinformation on social media and, by using a United States Senate bill as a guideline, I propose a bill that can be used in Canada. In the second part, I use the section 2(b) test to show how my proposed legislation might be found to violate the freedom of expression. In the third part, I use the section 1 Oakes test to show how my legislation would still be found as a justifiable limit in a free and democratic society.

Epidemics and pandemics are often described as unprecedented times; however, history demonstrates that these events are not as unique or uncharted as society believes. Several novels recount the experiences and impacts of historical infectious disease epidemics and pandemics around the world. Three such novels have been selected for discussion in this paper and will be contrasted with the ongoing news stories and accounts of the COVID-19 pandemic. This discussion will then cover almost 350 years of humanities’ relationship with infectious diseases, starting with the Plague epidemic in 1665 and extending to the continuing COVID-19 pandemic in December 2021.

The three novels selected each represent a turning point in the human-science relationship, and subsequently infectious disease. Daniel Defoe recounts the Plague epidemic in London in 1665 that killed a quarter of the city’s population in A Journal of the Plague Year. At this time, humanity possessed a scarce understanding of infectious disease and religion dominated discourse. The Ghost Map by Steven Johnson details the 10-day cholera epidemic in London in 1854 and the starting point for modern-day epidemiology and understanding infectious disease. The influenza pandemic that spanned from 1918-1920 is the focus of John M. Barry’s The Great Influenza. This pandemic is believed to have started in the army camps in the U.S. before spreading overseas and resulted in an estimated global death toll between 50 and 100 million people. The 1918 influenza pandemic represents the first intersection between infectious disease and modern-day science.

The novels demonstrate that commonalities exist across infectious epidemics and pandemics depicting what constitutes an effective, or poor, policy or legal response by government and public health officials. These commonalities exist despite the biological differences between infectious diseases and the evolving nature of modern science over time. Recognizing the patterns present in past epidemics and pandemics can help inform legal and policy decisions for effective official responses and preparations now and in the future.

Philip Barlow’s assertion that healthcare is not a human right is misleading. He assumes the right to health and the right to healthcare are interchangeable while discarding the critical role human rights play in public health work. Introducing the right to health into international law allows for maximum public health gains through state accountability. By enshrining a right to health in international law, states are held to a minimum standard in respecting, protecting and fulfilling the rights of individuals under their discretion. Health and human rights are inextricably linked. By recognizing that human rights belong in the realm of healthcare, officials can maximize their public health policies by considering the social determinants that impact an individual’s health. That said, the right to health goes beyond the right to healthcare, and it is crucial to continually distinguish between the two concepts.