AI Tool Sway Under Trial for Moderating Contentious Debates at Columbia University
Is it feasible to use artificial intelligence (AI) to moderate volatile discussions on subjects such as abortion, racism, immigration, or the Israeli-Palestinian conflict? This presumed potential is under examination at Columbia University. They’ve recently initiated a trial of Sway, an AI-enabled debating tool under development. Devised by a pair of Carnegie Mellon University academics, Sway aims to pair students with opposing viewpoints to engage in dialogues about controversial topics, in the hope of fostering superior interactions and exchanges.
Over 3,000 students across a network of over 30 higher education institutions have already tried out the software. The prospect of this collaboration arises amid growing tensions over the last two years within Columbia University involving students, university administrators, and federal authorities.
Sway orchestrates one-on-one debates between students with differing opinions, with an AI moderator present to enhance quality of discussion. Sway’s AI Guide in every chat aims to gently challenge students and push them to develop better reasoning skills through hard-hitting queries. Moreover, it purports to ‘suggest alternative phrasing’ if it finds the language used to be inappropriate or disrespectful.
The educator community at Columbia’s Teachers College is currently exploring Sway’s capabilities, considering its integration into systems meant for conflict resolution and to foster initiatives that create bridges in thought and understanding. Not only this, other teams within Columbia University have also shown an inclination to implement Sway in future academic semesters.
As part of an introductory video, Sway presented a simulated debate on a topical issue relating to whether the U.S. should focus more on Palestinian rights, ceasing weapon sales to Israel. This sample topic showcases the type of contentious conversations Sway could potentially moderate.
Columbia University is in the process of a major transformation in adherence to a $200 million settlement that is intended to address and reduce antisemitism within the campus. This settlement allows the university access to an extensive federal fund amounting to as much as $1.3 billion.
As per conditions of this hefty settlement, the university is required to maintain a substantial volume of data, implement rigorous rules against disruptive protests and improve further regulations involving international students. The university acknowledges that the employment of the Sway AI tool would contribute to the generation of constructive dialogues within the campus.
Despite promising signs, Sway’s approach has faced criticism, particularly from those who believe it seeks to resolve complex issues in a somewhat superficial manner. Detractors contend that applying such an approach tends to erase the deep historical and political contexts woven in these issues.
Critics also express concern that this AI tool might concentrate too much on mollifying participants’ feelings rather than addressing the real problems at hand. Doubts have also been expressed about the university’s approach to tackling these discussions, with one critic accusing Columbia of unsuccessfully searching for ‘magic bullet’ solutions to nuanced issues.
Despite this criticism, the developers of Sway assert that students are unquestionably aided by using the tool. Their goal is not necessarily to alter the students’ beliefs, but rather to inculcate a sense of openness toward opposing viewpoints with reduced animosity. In essence, the purpose is to foster civil conversation amid divergent perspectives, thereby encouraging understanding rather than agreement.