A team of four undergraduate students presented their research on ChatGPT and Introductory Computer Science Labs at USM’s twelfth annual Thinking Matters conference. The study, conducted by Devin Chamberlain, Abby Pitcairn, Nick Snow, and Ben Sweeney under the supervision of Professor David Levine, was presented in poster format.
Students already use AI tools to complete college work in many fields, sometimes appropriately and sometimes inappropriately. The team first created various personas seeking to embody how students currently use the tools. They then set about completing each lab in COS 170 and COS 161 using each persona. These various submissions were then compared to learn which labs (or portions of labs) were most susceptible or most resistant to being completed via AI tools. These results were then analyzed to determine common factors. The team’s work is summarized in three broad categories:
- Aspects of labs that make it challenging to detect students who are abusing AI tools,
- Aspects of labs that make it less desirable for students to engage with such tools and
- Opportunities to proactively include the use of such tools to augment learning.
The team is now generalizing the results by examining other AI tools and labs from other institutions.
“ChatGPT is a good place to start but a terrible place to end!”