Cedar Wins its First Artificial Intelligence (AI) Case

On May 25, 2023, the vice president of student services at Olympic College overturned the decision to issue a ChatGPT discipline sanction. Our client was a 16 year old Running Start student at Olympic College. He was accused by his professor of using ChatGPT to write a paper, immediately failed the course, and issued a disciplinary sanction.

In our client’s case, the professor ran the student’s paper through an AI-detection program and it reported that the paper was 75% AI-generated. As colleges and universities continue to navigate the emergence of ChatGPT, AI-detectiontools have sprung up to meet increasing demand for a technology capable of determining whether a piece of written text was AI-generated or human-generated. AI-detection programs check for certain linguistic features or patterns that are common in machine-generated text, assessing the formatting and structural features such as repetition, unusual word choice, rhythm, and variability of text to make a prediction as to whether the text may have been developed by a machine.

The critical distinction between AI-detection programs and standard plagiarism checkers is that AI detection programs are not scanning the web for duplicate material and simply reporting a percentage match to a text entry, but instead using predictive metrics and algorithmic tools to make an educated prediction of machine-generation. As a result, the AI detectors currently available cannot reliably detect AI-generated text. In fact, the VP’s own email to us acknowledging receipt of our appeal notice was reported at 66% AI-generated!

Not only do AI detection programs regularly provide false positives, but they are also biased against non-native speakers – a recent study showed GPT detectors were misclassifying writing by non-native English speakers as AI-generated 48-76% of the time, compared to much lower figures for native speakers.

Cedar understands that Olympic College was faced with an extremely challenging dilemma with respect to emerging AI technologies and the ever important need to honor and promote academic integrity amongst its students. Many higher education institutions across the country are grappling with this complex issue in ways that simultaneously uphold foundational educational principles while also honoring students and their rights to fair, honest, and transparent adjudication of disciplinary matters. The lesson from these cases is that colleges must be extremely conscientious with respect to the present lack of AI-detection tools and the importance of evaluating evidence before them to arrive at a just result.

Published by: Luke Hackenberg and Rae Podrebarac