This Saturday, the Jefferson Literary and Debating Society held its first-ever Martin Skeptical Debate in Old Cabell Hall. The Martin Skeptical Debate, which will be held annually, is the product of a grant made to the Jefferson Society in the will of the late R. Bruce Martin, who taught chemistry at the University for 35 years. The purpose of the debate, as stipulated by Martin, is to promote scientific skepticism at the University. The resolution for the inaugural debate was as follows: “AI will significantly increase biological risk in the next 5 years.”
Ryan Shoztic, President of the Jefferson Society, began the event with an opening address on the background behind Professor Martin’s Grant and the relevance of skepticism and debate to the University. An address from Daniel Remler, a former diplomat and senior fellow at the Center for a New American Security, followed Shoztic’s speech. Remler described his own experience with artificial intelligence policy as a diplomat, noting in particular the current efforts by the People’s Republic of China to integrate artificial intelligence into their own diplomacy, government operations, and military infrastructure. He also noted two schools of thought regarding artificial intelligence: one focused on restricting it, and one focused on letting it develop unimpeded. He emphasized the downsides of both positions, regarding them as contrary to the principles of liberalism. After Remler, Amelia Cantwell, Debate and Oratory Chair of the Jefferson Society, thanked the panelists and judges for attending and read out the rules of debate.
There were two teams in the debate, each with two members. The affirmative side consisted of Varnum Bayless, a first-year student at the UVA School of Law, and Christoph Schoer, a fourth-year student in economics and politics. The negative side consisted of Lisa Berlizova, a fourth-year computer engineering student, and Jason Chin, another fourth-year student in computer engineering. The contestants were judged by Remler, as mentioned above; Dr. Adeline Williams, a biological scientist at the RAND Corporation; and Dr. Kelly Seagraves, a neuroscientist and Vice President for National Security and International Affairs at the Biotechnology Innovation Organization.
Schoer began with the affirmative’s case. Schoer argued that the biological risk of AI was tied to its potential for usage in agro-terrorism, in addition to general concerns about the possibility of AI being employed for the ends of other forms of bio-terrorism. A central component of this contention is the fact that artificial intelligence has been used to generate models of certain proteins, bacteria, and other biological structures at speeds that had previously been unattainable, thus allowing terrorist organizations to engage in agro- and bio-terrorism with far less effort than previously possible.
Berlizova began with the negative case. She argued that AI has provided significant benefits in the medical field in the past few years. Among the most prominent contributions made by AI in recent years was its role in the development of COVID-19 vaccines five years ago. She also made the argument that artificial intelligence can help democratize access to medical technology, as using it does not necessarily require specific expertise in the medical field.
After cross-examinations on each side, the teams moved into rebutting each other’s arguments and furnishing their own with further evidence. Bayless offered a critique, for example, of the negative argument surrounding the role of AI in developing the COVID vaccine. In his view, the concentrated efforts of world governments were more relevant to this process. He also emphasized the risk of data distortion when information is put into large language models, potentially giving doctors inaccurate information. Chin offered a rebuttal to the affirmative’s points regarding the role of AI in creating a COVID vaccine, citing evidence from pharmaceutical companies on the use of the technology. He also made the case that the affirmative was misunderstanding the capacity of terrorist groups to make use of AI and exaggerating the present threat.
Much of the final part of the debate dealt with the capacity of people to use AI, for better or for worse, the extent to which AI will actually advance in the next few years, and the extent to which we can attribute modern medical developments to AI. Both teams concluded with arguments emphasizing the standard for what counts as “increasing” biological risk, specifically, regardless of other benefits of AI, and criticizing the opposing side’s understanding of the “next 5 years” specification in the prompt, respectively.
Following their deliberations, the judges eventually reached a verdict, and Seagraves went up to announce the winners. She gave the affirmative the upper hand on oratory and the negative the upper hand on command of evidence, but ultimately, based on their ability to effectively tie their core argument to the prompt, gave the victory to the affirmative.
Leave a Reply