AI Policy at UCLA: No TurnitIn Detector - Part 2

In a prior post, we noted that official UCLA policy is not to use TurnitIn's supposed detector of AI-generated content in student writing.* The Academic Senate indicated that instructors should announce that students shouldn't submit AI-generated content, but it did not get into how such work might be detected.**

While an essay that is grammatically perfect but contains factual errors is suggestive of AI content generation, use of AI can't be proved.

Inside Higher Ed today has a story about an instructor at another university who inputted student papers into chatGPT and asked it if the essays were AI-generated. He then sanctioned many students who denied using AI.*** The problem with this method is that chatGPT, just as it is capable of making errors or seeming to make stuff up, cannot be relied on as an AI detector.

Yesterday we posted a video from a segment of a Regents meeting on AI in which a Berkeley expert slammed those telling sci-fi stories about AI potentially becoming sentient and taking over the world.**** However, apart from saying that the usual way of teaching writing was no longer going to work, his suggestion for dealing with that development was limited. He suggested presenting students with an AI-generated essay and asking them to critique it. It's not clear that such an approach is a full substitute for learning basic writing skills. The sad fact is that other than telling students not to submit AI-generated content, we don't yet have a simple answer to the problem.

===

*https://uclafacultyassociation.blogspot.com/2023/04/ai-policy-at-ucla-no-turnitin-detector.html.

**https://uclafacultyassociation.blogspot.com/2023/03/ucla-senate-guidelines-regarding-ai.html.

***https://www.insidehighered.com/news/quick-takes/2023/05/19/professor-students-chatgpt-told-me-fail-you.

****If you didn't see it in our earlier post, here it is again:

Or direct to https://www.youtube.com/watch?v=AygJUbwCwJc.