Samantha Newell from the University of Adelaide has extensively tested the limits and output of ChatGPT to help identify constructed submissions.