Students Debate AI Governance at Waterloo AI Forum

By Auz Momin

Earlier this month the Waterloo AI Forum Debate was organized by a student-led initiative called Effective Policy. There were spots for 40 students to share their perspectives on AI with an opportunity to take home a cash prize. Debate motions included AI taxes against unemployment, which military uses of AI should be permitted, how to deal with electoral deepfakes, misinformation and more.  Branka Marijan, a Senior Researcher at Project Ploughshares and a participant of the Centre for Peace Advancement, was one of the judges at this debate.  

The winners of the debate were Alexander Horoyski (Honours Science) and Finn O'Connor (Political Science). They debated whether object detection technology used to identify soldiers, artillery, and military vehicles should be licensed to US allies to install on their own drones. Their arguments to not allow this were:  

1. Although autonomous weapons may reduce casualties in any one war, they may drive arms races that lead to more total wars (which doesn't necessarily mean fewer casualties overall) 

2. Allies can change as new leaders come to power. The proliferation of autonomous weapons leads to little control over how they might be used in the future.  

3. AI systems are inherently unreliable when the data input to them changes. We can't predict how autonomous weapons will behave in novel environments (ex: inter-urban battlegrounds) 

UWaterloo Effective Policy aims to address rapid change in the world. They believe that “current governance decisions will shape the rest of our lives and we should help improve them.” Effective Policy works by connecting students and professors who care about a specific policy problem so that together they can develop better policies for their problem. “AI is moving so fast that we don't have existing playbooks on how to precisely govern it. It isn't the case that students or ordinary citizens are lacking some insight that academics and researchers have.” explained Madav Malhorta, the founder of Effective Policy. 

Madav is a computer engineering student at the University of Waterloo. He has always been researching how to mitigate social problems like AI-enabled misinformation and enhance safe AI applications through his ventures. Madav founded Effective Policy because he “found it hard to relate to students in my bubble in computer engineering who were focused only on getting big tech jobs.” This led to Madav spending time with students, professors, and clubs across all faculties at the university. “I saw that every faculty had its bubble of students who were interested in doing greater good, but these students weren't collaborating together or benefiting from each other's interdisciplinary expertise.” explained Madav. By launching Effective Policy, he has created a platform for students to “collaborate on solving important social problems.” 

Effective Policy has received support from multiple organizations and professors, including Centre for Peace Advancement Director Paul Heidebrecht. There is a clear alignment with the Centre’s efforts to facilitate interdisciplinary learning through opportunities such as the Map the System competition and PeaceTech Living Learning Community. “We're very grateful to all the professors for their continued mentorship.” said Madav. Effective Policy will be recruiting a new cohort of students to pair with their network of mentors in May. Interested students can find more information on their website.