Within months of the release of ChatGPT in November of 2022, the first AI detector appeared on the scene. Since then, new AI detectors have been released almost weekly, each claiming to be able to identify AI-produced or edited text. Yet in the scholarly commentary on AI, AI detectors (themselves a form of AI) are routinely criticized for their inaccuracies, inconsistencies, and high number of false positives. So what’s a teacher to do? Do they work or not?

In this presentation, Andrew Higgins will try and walk you through the world of AI detectors, explaining how they work, their uses, and their limitations. He’ll argue that, while AI detectors can’t do all the wonderful things they claim they can do, they are still useful tools for teachers trying to ensure that students are actually writing the work they’re submitting to us. In the process, he will share data he has collected on how several prominent AI detectors respond to texts written before and after the emergence of generative AI. We’ll end by talking about the future of AI detection and sharing strategies people have for handling this complicated issue.

Come join us on Wednesday, December 4th, at 2:00pm in the FDC or on WebEx!