Reflecting on the First 2 Weeks of AI-(Em)Powered Instruction

Updated Post:

It has been an incredible two weeks in the classroom. I have seen students' feeling true excitement about learning, I have heard their stories of feeling empowered (as you (hopefully) can hear Madison discussing about her business) in areas they once found extremely intimidating (i.e. at a car dealership); students are going home and are WILLFULLY LEARNING about things they care about, they are seeking out different perspectives and considering information otherwise impossible for them to have known existed. It's incredible. Honestly, if that doesn't show a more productive citizen in the making, I don't know what does.

My First Year Teaching (Again)

As for the teaching of AI, it is definitely a new and intimidating concept to introduce to students. I have been incredibly honest with them about not having all the answers and they (not surprising) have shown amazing kindness and patience. If I have learned anything over the past two weeks, it's that I will be reteaching most classes at least once to become more proficient at teaching prompt engineering. It's weird, I almost feel like I'm back in my first year of teaching; I am very much still learning how to navigate the AI terrain myself, while also trying to guide students down the easiest path, knowing full well I have no idea what the path looks like or if it even exists yet. Anyways, it's all very exciting.

Big Takeaways

Here are some big takeaways I've been dying to share:

- Framing the chatbot as a *tool* is essential to student's being respectful of its purpose. I used the following analogy: "If you go to hire a construction team and you see some workers pounding everything with a hammer, but others having the awareness of the other tools available to them that would be more appropriate for the task, you know who's going to end up getting the job." My main goal here is to let students see that someone who doesn't know how to use a tool well (or when to use it, or how it works, etc.) is always going to be seen in a better light than someone who doesn't. No one wants to be the person who's full of hot air, know what I mean?

- Another eye-opening moment for me was the level of writing skill it takes for students to be *good* prompt engineers. It's essentially asking them to do the EXACT same thing I would teach them pre-AI: how to write effectively using focus, purpose, audience, organization, and development to inform their writing process. The students realized quickly that putting together a *good* prompt (one where they got the result they wanted, in the way they wanted it, etc.) was not only hard work, but next to impossible. Organically, students came to the realization (thank the heavens above) that chatbots will never make the result the student has in their minds' eye because it is just a tool. Also, their brains did start to melt after about an hour of practicing (and so did mine).

- If you are attempting to roll out AI with students who are brand new to chatbots, be sure to provide multiple options for how to frame prompts (I used a RASCEF as the "simple" framework, the 5 "S" model for the "mid," and an extensively developed RODES sample prompt for the "complex"). Again, I do teach a co-req model for students who did not satisfy Texas's reading and writing college level proficiency exam (otherwise known as the TSI), but it's been my personal experience that most students are not inherently strong writers. Giving them the ability to scaffold their prompts into bite sized writing chunks definitely helped.

Original Post:

This week was an incredible learning experience for both my students and myself. I will be updating this post later this week, but I wanted to make sure everyone has access to real student voices, like those of Madison and Mo (linked below). I plan to continue documenting student experiences via discussion posts and interviews as the semester progresses.