I posted a new rubric on my LinkedIn this morning, basing it upon the list that Eugenia Novokshanova and I had developed in the summer of 2023 as we were writing our first writing course with AI. We wanted to define what skills our students would need in five years as they worked for an employer in the Age of AI. These were the skills we identified:
✔️ They can follow directions.
✔️ They write with a unique voice, appropriate tone, and demonstrate an understanding of their audience's needs.
✔️ They are transparent about their AI use.
✔️ They are careful about facts, and reference facts correctly and completely.
✔️ They demonstrate proficiency with AI tools and use them efficiently and ethically.
In my post, I followed that list up with, “I hope to carry those skills through all my rubrics this year. I want students to carry their understanding of ethical, responsible, and efficient use of AI through everything they do.” Then I posted my new rubric that was based on those skills:
I was feeling pretty smug about my posting, thinking how this list had served me so well for going on two years, when, unexpectedly, came the very astute comment of Emily Pacheco, founder of EDHUB.AI:
Helpful rubric! One additional one might be “can they solve a problem?” I want people who can help solve problems, not just follow directions!
“OHMYGOODNESS!” I thought to myself. “How could I have ignored such a practical and obvious skill?”
I had just presented about the importance of students “making trouble” with AI—learning to push back to make their own voice and academic intent a priority—but I had never even considered problem solving in a focused way, especially in my rubrics. What was wrong with me? My entire class is based on problem-solving, with project-based and problem-based learning at the core, but I had not been explicit about my problem-solving in my class. It was more like . . . a given? (When would I learn that there are no givens!)
UGH.
As I sat and thought about what Emily had so helpfully offered, I decided two things:
It was definitely time to reconsider that original list of skills.
I needed to directly address problem-solving (with and without AI) in my classes. I need to think about this and research it, so I’m going to address this in a later post.
Reconsidering the List of Skills
So, the first thing I needed to work on was that aging list. The list had worked great at the beginning of my teaching with AI, but now I had a much better understanding of the challenges AI presented to me and to my students in an educational environment. It was definitely time to expand and reconsider priorities!
Following Directions
I started out thinking I was definitely not abandoning “They can follow directions,” because this skill involves two things that most of my students sorely lack—(1) the ability to read and comprehend instructions, and (2) the humility to understand that others might be as smart (or smarter) than they are and may have a clear idea of what needs to be accomplished. But maybe I was getting too curmudgeonly in my role as a professor. Maybe I needed to add a little more context. How about “Follow directions, analyze outcomes, and adapt to change”? I guess I could put that addition under a different skill, but it seems to fit here. Implicit in this grouping is the idea that they would first follow the directions, then analyze the outcome, then adapt. But what if the directions were very good and didn’t need changing? Hmmm. Maybe I needed a few more words: “Follow directions, analyze outcomes, and adapt to change when needed.” That would connote that they didn’t have to be in a constant state of change, but they could be flexible if the situation called for it. It would also assume that they had the know-how to understand when change was needed and when it was not (I know, I know, that’s a big ask!)
Voice, Tone, and Audience Awareness
The next one, “They write with a unique voice, appropriate tone, and demonstrate an understanding of their audience's needs” seemed pretty cut and dried, but there is one thing I really want to add here: the idea that I don’t really care if they use AI, I just care that they don’t sound like they use AI. This list of skills is essential to capturing a human sound in writing. How about, “The ability to write or edit AI to capture a unique voice and appropriate tone in sync with an audience’s needs”? I want to get away from the idea that any AI use is somehow “less human.” The output of AI is human-derived and human-prompted. There is no reason why revision of that content isn’t considered to be as “human” as any other written artifact.
Transparency
I might be getting myself into a little hot water here, but the idea of being “transparent about AI use” is something I am starting to question. Yes, transparency in education feels right, as we want our students to be keenly aware of how they are using AI, what AI they are using, and why they are using it, but I also strongly feel that the idea of transparency should only be applied in a classroom where students are learning to use AI in a supportive environment. Outside of that supportive classroom, I think it is a trap. I have already warned my students about supplying a transparency statement in a class where the professor is openly hostile to AI use. After all, if they are using AI responsibly and ethically, there really is no reason to disclose that use, and if they are using it unethically, they will never disclose it anyway! In the world of work, a transparency statement seems weird. Maybe some publishers might like one, but let’s let them ask for it. I think employers, to a large degree, won’t want transparency in the future—they will just want writing to sound plausibly human.
I will continue to require a transparency statement from my students because they are learning to use AI in a supportive environment, but I will definitely asterisk the use of transparency statements outside of my classroom.
When there is a witch-hunt going on, one should never concede to being a witch.
Facts
The thing about facts is that one has to have content knowledge to know whether facts are or are not correct. Without content knowledge, one may not correctly understand the context of the facts that are presented in addition to the facts themselves. The idea of “fact-checking,” while significantly important to protect the veracity of a piece of writing, can’t replace content knowledge. It is imperative, therefore, that one does not write ANYTHING with AI unless they first possess content knowledge of the subject. So, I think I will revise this skill to read, “Deep understanding of one or more content areas of a particular profession, business, or industry.” Thorough research, fact-checking, and revision are not always enough to be factual, and we need to ensure that employers are aware of the need to hire competent, knowledgeable humans before attempting to use AI. Content knowledge provides context of a subject and makes it possible to thoroughly and completely fact-check.
Effectiveness
Effective use of AI takes time and effort. One must first understand the scope of tools available for a particular job, and then they must understand how to use those tools. For one to be effective with AI, they must be lifelong learners, working daily to keep abreast of new tools and trends. They must be curious about new use cases, and they must experiment on their own. “They must have a strong commitment to exploration, a flexible mindset, and a broad understanding of AI literacy.”
New List Items
“They must be resilient and critical thinkers, ready to question results and demand better answers.” If you are going to work with AI, especially in a profession, you must be willing to push against it when it doesn’t give you the best answer or the most thorough answer. This requires critical thought and analysis of AI output and an understanding of what questions to ask and how to ask them.
They must be problem solvers. This requires creativity, research, and an encyclopedic knowledge of the content knowledge required to do a job. How can you design a solution if you don’t know the context, the library of tools available, and the possibilities before you? A problem solver is a person who doesn’t back down to a challenge, who pushes the envelope, and who finds a way around, over, or through a problem. These will be the highest-paid members of the Age of AI, the thought leaders and entrepreneurs of tomorrow.
OK. So here’s the new list of essential skills I think my students will need when they are employed to work with AI five years from now:
They can follow directions, analyze outcomes, and adapt to change when needed.
They can write or edit AI to capture a unique voice and appropriate tone in sync with an audience’s needs
They have a deep understanding of one or more content areas of a particular profession, business, or industry, so they can easily identify factual errors.
They have a strong commitment to exploration, a flexible mindset, and a broad understanding of AI literacy.
They are resilient and critical thinkers, ready to question results and demand better answers.
They are problem solvers.
And, of course, here is a new rubric built on those skills:
Michelle, thanks for this. It is a great demonstration of unpacking what new LOs we should be writing into our courses and units. Content knowledge is dead. Skills, especially critical thinking and problem solving, are now the main game.