ChatGPT Writing Curriculum Notes

This semester, at National Taiwan University, I’ll be teaching “AI-Powered English Writing Application & Advancement,” a course for teaching general academic writing and using LLM+other AI tools to help with that, with a particular focus on ChatGPT. The students from this class are from all levels and all departments. My assignments will almost entirely be meta in nature, asking students to write about the use and future of LLMs with regards to education, workplaces, and their fields.

I’m compiling a list of links I’ve used to help me plan the course in this post. Of course it is not complete, but it might be a good set of starting places for someone else trying to accomplish something similar.

Definitely interested in more links if they are meaningful.

Right now not in a particular order, might revisit later.

University-related Resources

MIT Prompt Engineering Coursera Course — This is a six week course. The first three weeks of material have a lot of good suggestions for creative ways to approach giving ChatGPT prompts. The last three weeks did not feel as useful, nor did the assignments. Also, one negative about this course is it builds off the assumption that you can generally trust what ChatGPT tells you informationally, which is not good. The course also featured three arxiv papers ( 1 , 2 , 3 ) which I have not yet read (they’re more tangential to the curriculum I’m aiming at).

UT Austin Center for Teaching and Learning — Resources for instructors, with many many links.

UT Austin student paper The Daily Texan — Well-written student opinion piece arguing for the usage and teaching of ChatGPT.

Ohio University Center for Teaching, Learning, and Assessment — Looks like a big collection of assignments using ChatGPT, unfortunately they’re behind a SharePoint login.

Wisconsin University Writing Center — Really great collection ideas and links for using ChatGPT in academic writing, including the negatives (plagiarism, IP issues, etc.).

Massive crowd-sourced collection of policies around LLM usage .

Popular Media

New Yorker — Negative longform article on the purpose of ChatGPT.

Writers’ Striker — Funny image for lecture purposes.

Twitter thread — Stories about professors being unable to understand ChatGPT is not really detectable.

Ars Technica — Why detectors don’t work.

The Verge — Lawyer using ChatGPT-hallucinated cases.

The Guardian — Vanderbilt used ChatGPT to respond to a school shooting. I think this is a good discussion starter, since IMO this is a situation where ChatGPT is very useful (e.g. an email needs to be sent that you probably have no idea how to actually write and is probably not going to be meaningful no matter what you say).

Other LLM/AI-Powered Tools

Note — I have not tested these beyond looking at them, so I’m copying their abouts.

Paper Digest — “An AI-powered platform to track, search, review and rewrite scientific literature.”

Elicit — “Elicit uses language models to help you automate research workflows, like parts of literature review. Elicit can find relevant papers without perfect keyword match, summarize takeaways from the paper specific to your question, and extract key information from the papers. While answering questions with research is the main focus of Elicit, there are also other research tasks that help with brainstorming, summarization, and text classification.”

Litmaps — “Litmaps is the easiest, quickest and most accurate way to find the articles and papers you need to complete a literature review.‍ From a single paper, Litmaps generates a map of the most relevant articles that relate to your seed paper. The most recent articles appear on the right, the most cited articles appear at the top and the lines show the citations in-between.”

Scite Assistant — “Assistant is a conversational tool made by scite that lets you ask questions in simple language and get answers backed by real, up to date references. Think of like ChatGPT with real, up to date references, tailor-made for anyone discovering, understanding, or writing research.”