Overview of the Cheat-a-thon
Penn State’s Center for Socially Responsible Artificial Intelligence recently held its inaugural “Cheat-a-thon,” a virtual competition aimed at exploring the capabilities and limitations of generative AI in an academic setting. The event ran from March 3 to April 6 and invited faculty and students from various U.S. universities to participate. Faculty submitted challenging exam questions, while students provided AI-generated answers. The competition sought to shed light on how well AI can handle college-level queries.
Key Details from the Competition
- Faculty from ten universities contributed 66 difficult questions spanning 18 disciplines, such as accounting, physics, and humanities.
- A total of 33 students submitted 451 AI-generated answers to these questions, showcasing the AI’s performance.
- Five faculty members received $1,000 prizes for the most challenging questions, with winners coming from Penn State University and the University of Pittsburgh.
- Student winners, all from Penn State, were awarded prizes for their high scores, with one student achieving a perfect score across multiple subjects.
Significance of the Findings
This competition highlights the growing concern over academic integrity in the age of AI. Many educators fear that students might misuse AI tools for cheating. By conducting the Cheat-a-thon, Penn State aims to understand AI’s impact on education better and enable faculty to create more effective assessments. The insights gained will help educators adapt to the evolving landscape of academic integrity, ensuring that future exams and assignments remain robust against AI-generated responses.
