Sometimes I'm called upon to teach a writing intensive capstone class where the main assignment has been a review paper. Given #GenerativeAI, I've been wondering what to do differently. Helping students improve their writing is totally different now...that's all I know.
I found this article:
The role of ChatGPT in scientific communication: writing better scientific review articles
@rspfau my feeling is we're simply going to have to move to different forms of assessments.
Either written assignments where real time internet access is denied (I.e. No #generativeAI available) or, my preference, the need to combine writing tasks with short presentations and Q&A. You can't fake understanding of a topic in public, and will quickly work out who's done their own writing or not. @academicchatter#academicchatter
@gpollara@eilonwy@rspfau@academicchatter There’s a Stanford professor who’s simply integrating LLMs into his lesson plan - he accepts work generated by LLMs, but students must include the prompt used to generate the text and are graded on that as well.
Seems to me adapting to the reality of LLMs and teaching students to use it effectively is a more successful approach to teaching.
@dogzilla@eilonwy@rspfau@academicchatter to be fair, I don't dislike that - teaching is not just about passing on knowledge, it's also about preparing students for the next steps in life.
Whether we like it or not - #generativeAI will be with us going forward, so might as well teach how to extract its benefits - e.g. examples of how to build and use #ChatGPT bots, etc... #academicchatter#ai
Add comment