Maybe it is a failing of mine, but feedback takes me awhile. I think this because I care very deeply that it be helpful for the students. I agonize over it... So, yeah, this saves me a ton of time, is my authentic feedback, and is customized to the individual.

Crafting a prompt is a form of discourse, a new literacy if you will that relies on the literacies that came before. Reviewing and revising the material and AI generates in response to a prompt is an act of critical thinking and critical discourse.

I guess as a Multiliteracies scholar who's been working in the field for 14 years and gone through extensive research, peer review, and publishing, I have a different perspective on discourse. To me, a recording of me speaking is a form of discourse, a form of text that has the same value prose.

So, I had to go into the longer transcript and separate each one out to ensure the individualization. It works great now. It only includes the feedback I provide and organizes that feedback the way I prompt it to.

I create a separate recording for each student. This way each transcript (and subsequently each AI synthesis of that transcript) focuses on the individual and their needs. The first time I did this I had a long recording and it did homogenize the feedback across students as described in the article.

I always go back to the concept of remixing as described by the New London Group. Everything is iterative becoming new texts as we adapt and change them. New and Multiliteracies scholars have been thinking about these processes since the 90s. www.sfu.ca/~decaste/new...

You're presuming the software is the one prioritizing. When I use this process, I direct it to organize and prioritize in my prompting and then, revise a fairly good amount. I've written and published plenty using many different processes for many different outlets. This manner still feels like me.

Think about it this way: We would allow a student without the ability to type/write audio record their idea and turn that into a draft. Is this process really that much different? What is considered an accommodation is really just a supportive practice that benefits all students.

Agreed. I'm stealing that one for sure!

I was referring to LLMs, since that was the focus of the thread. I've found:
Copilot 👌🏻
Perplexity 👍🏻
Get Liner 🙌🏻

Curious about ChatGPT. I haven't tried it yet. It sounds like in your experience it hallucinates content. Is it topically way off, or is more like it presents intro/lit rev as findings?

I'm jealous of your educating experience. Elementary kids using Boolean operators in research sounds amazing. My high schoolers rarely know what they or how to use them effectively. We get them there, but it takes time and can be intimidating. So having alternatives isn't a bad thing.

I didn't mean to imply that Google Scholar required Boolean operators. Other than formal databases, the search engines don't require them. Boolean operators are great for databases but not as effective in web search engines (Google, Safari, etc.).

AI doesn't require Boolean operators to conduct a search, which can be helpful for students unfamiliar or uncomfortable with the process. That said, I would recommend students use AI early in a search to find a few resources to get them started or when they feel like they aren't finding "anything".

I haven't found search engines super helpful for finding peer reviewed research articles, except maybe Google Scholar, in part because there's just so much information on the Internet. Even Boolean operators can fall short when it comes to web-based search engines.