The seams had weakened with use, and I stood next to the car, hands full of things I needed for the run, my gear outnumbering the pockets in my shorts. Thankfully, my running partner had pockets, and we sorted it out, but I was surprised how much it up-ended me, changed my mood, and even made me wonder if I should continue with the run.

Yes, friends. Another metaphor. Here’s a story that connects running to my AI use.

I handed my thinking to AI and asked it to build something; it broke at the weak spots.

The Stretched Out Seam

I work in teacher preparation, which means an endless list of what’s due–for them and for me– what’s coming up, who needs what, when. It often feels frantic, so I asked Claude to help build a workspace for the juggling balls: a place to track class preparation steps, tasks, deadlines.

I used it up! I added tasks by hand. I dumped raw thinking into the platform it built— the kind of half-formed notes you only write down because you trust you’ll come back to them. Then I crumpled up the notes and threw them away.

I was giddy with executive function bliss.

I loved it so much that I asked Claude to make some revisions, a few changes, but I wasn’t specific about saving what I had added to the list, lured by the little plus sign that was keeping track of my brain. So Claude, bless its heart, built a new version of the platform, generated fresh from the conversation, not layered onto the one I’d already been filling in by hand.

And the little darling didn’t save a thing.

The manual additions were gone. (!) After regulating my heart rate, I dug sticky notes out of the trash. I remembered some tasks over the course of the day, out of order, out of context. Some of it — the thinking that went straight from my head into that workspace and nowhere else — is probably just gone.

The truth: I was overwhelmed. Scattered. Busting at the to-do list seams. I didn’t build a tool with a clear picture of what I needed. Instead, I hoped the tool would organize me. I wanted the overwhelm to stop, fast, and a request made from that place is a request that skips a crucial step: deciding, before I said anything, why I was building the thing in the first place!

Efficiency tools are supposed to save the thinking you don’t need to redo. Tools can’t help if the human’s directions are flawed. This was my weak seam.

I’m back to a pen and paper planning until I get settled into the semester, but I hope to work on a fix and try the build again. I need more than better directions; I need a habit of stopping, for even ten seconds, to make sure the human knows what she wants!

With that said, better language still matters, turning the pause into prompting language. I asked Claude how to improve my directions. It noted a few options:

“I’m asking you to add to this, not rebuild it. If anything I say conflicts with what’s already here, stop and ask before you overwrite it.”

“Some of what’s in here only exists in this document — I haven’t saved it anywhere else. Add what I’m asking for now without regenerating the rest, and don’t remove or replace anything without checking with me first.”

“Before you make any changes, show me what’s staying, what’s new, and what would be removed — so I can catch anything I’d lose before it’s gone.”

That third one is the one I’d pick because it forces a checkpoint. It makes the machine show its work instead of trusting that it read my intent correctly, or that I knew what I wanted without some built-in reflection. The pause can be in my head and be built into the request.

The Lesson

I offloaded my thinking and asked AI to build something from my idea, which was innovative and responsive (ok, maybe a little too responsive). I reached for a quick fix instead of pausing to decide what I wanted first.

Efficiency backfires the moment it replaces thinking instead of following critical judgement and detailed directions. The humans are still in charge here, even when we mess it up. A machine can move fast, but it can’t know what I needed to protect if I didn’t stop to know my needs.

The human move — the one that’s easy to skip when you’re too busy to think you have time for it — is to pause inside the innovation loop. Not after the output comes back, but before you ask, and again while you’re reading: is this what I actually meant, or did I just get an answer fast enough that it felt like relief?

One Last Thing!

The habit I want to foster underneath all of this, and it’s not really about AI:

Know where your seams are.

Know what you can’t afford to lose.

When you’re the most tempted to skip the pause because you’re overwhelmed, because you just want it handled—that’s exactly when it matters most.

I’m still working on the first one. ;)

Still learning,

Serena

I’m Serena Morales, Clinical Associate Professor at Boise State University, AI professional learning facilitator, and an educator who believes pedagogy is central to any conversation about education. PromptWise is where I think out loud about what it means to develop AI-integrated pedagogies in learning settings.

What’s new

A lot has happened in a year! Alongside teaching, I spent the year writing PromptWise every Thursday, facilitating AI professional learning with K-12 teachers across Idaho and university faculty at Boise State, and presenting at local and national conferences. I also finished something I’ve been building — Unprompted: Designing Assessment in the Age of AI, a free book you can download, use, remix, and reuse.

The landscape shifted, too. Idaho passed SB 1227 this spring, requiring every district to adopt an AI framework, develop AI literacy standards for students, and build professional development for educators. AI integration is no longer optional, which means what we’ve been doing here, and what you are doing there are slowly moving towards each other.

PromptWise is here to help! The newsletter will post every other week with the same pedagogy-first format. Each issue gives you a prompt and a practical example, plus something new: a pedagogy move—one named instructional strategy—that connects the prompt to what you already know about good teaching and learning.

Some issues will go bigger: a chained sequence of prompts that build on each other, or a live tool you can try directly with students. And again this fall, I’m opening PromptWise to guest writers. If you’ve tried something, built something, or learned something the hard way, I want to hear from you!

Where we’re going

In September we’re moving into instruction and thinking about what it looks like to leverage AI to build teaching experiences, to put ideas into action. Maybe you want to build a student practice tool that responds, adapts, and gives you data before the test. Or a feedback tool that asks a follow-up question. Perhaps you have an idea for a learning experience that is worth the device time, especially now, when device time is reduced in many classrooms.

We’ll also explore how AI continues to develop as a competent assistant that saves time you can reinvest with students, in energy for your life outside the classroom. I’m not talking about offloading what it means to be a teacher; I mean protecting why you are an educator by offloading tasks that cause burnout.

One Last Thing!

During a workshop I facilitated last year, I overheard a conversation (about me! woot!) from the back of the room as folks were getting settled after a break. It went like this:

Person A: Well of course she’s going to say that. She loves AI.

Person B: I don’t know. It’s all just hard.

Person C. [shaking head, forlorn?]

I pulled back into the hallway, took a breath, then approached the group. I told them I had heard them chatting, and that I wanted a chance to clarify my stance on AI. Now seems like a good time to repeat it:

How I feel about AI doesn’t really matter. Some days it’s my bestie, my assistant, my thinking partner, my scheduler. Other days it wrecks a system I thought I had, gets dates wrong, repeats itself. On all days, it’s a tool—like any other technology humans wrangle over time—I feel obligated to understand, to learn, and to leverage.

The through-line for everything (I wish you could see my arms spread wide in emphasis as I say this): Pedagogy first! AI somewhere after. Same as always.

Still learning,

Serena

PromptWise is a free newsletter for educators exploring generative AI. If this was forwarded to you, please subscribe. Interested in professional learning for your school or district? Reply or reach out at serena.promptwise@gmail.com.

It’s a framework for thinking–starting with your own practice–about how to protect learning and augment pedagogy in a changing education environment.

Click on the link to download or print the full, free book.

Here’s what’s inside:

Chapter 1 reframes four assessment principles you already know — validity, reliability, fairness, and equity — through the lens of AI, then adds a fifth: Explainability. Can students account for their own thinking? How does explaining learning make it visible? And human!

Chapter 2 takes tried-and-true frameworks (Backwards Design, Bloom’s Taxonomy, the AI Assessment Continuum) and adds a new filter question: Where does evidence of learning live in this assignment, and is that placement intentional?

Chapter 3 is about designing work that AI can’t do without the student by building tasks around thinking, experience, and voice that are specific to each learner. Visible Verbs. Personal Anchoring. Entry Fee strategies. The tools are concrete and the moves are transferable.

Chapter 4 introduces the 10-80-10 Workflow, Friction Filters, and Take It. Leave It. Learn It. Practical, usable, adaptable to any grade level or discipline.

Chapter 5 gives you language to lead conversations about AI with students, colleagues, administrators, and families.

Who is Unprompted for?

Classroom teachers who aren’t sure how to integrate AI. Instructional coaches looking for a shared framework to bring to their teams. Teacher educators who understand the urgency of preparing AI-ready teachers. Curriculum directors and deans who need practical resources. Administrators who want to move their staff beyond policy and into practice.

If you design learning experiences at any level, in any discipline, this book was written for you.

How can you use it?

Read it! and apply the frameworks to your own practice. Assign a chapter in a graduate course or a PD session. Use the lesson plans in Chapter 5 directly with students or staff. Adapt the tools for your context.

It’s free. It’s openly licensed. Read it, share it, adapt it, hand it to someone who needs it.

AI changed up the teaching landscape, but the expert in the room is still you.

Learning like crazy,

Serena

Most educators I work with have plenty of information about AI and lists of tools that are allowed or restricted. Now they need structured time to work through the hard design questions about instruction, assessment, and disciplinary applications with colleagues.

I can’t seem to keep my face out of the information: As of December 2024, 31% of U.S. public schools have a written AI policy, and 68% of teachers have received no AI training from their school or district. This is a story about the distance between state-level guidance and local implementation. When guidance is advisory rather than mandatory, adoption tends to be uneven, and the districts with the fewest resources to figure it out on their own are often the ones that need support most. That inequity is already showing up in how students are accessing and using AI. States are setting ambitious expectations. The professional learning infrastructure to meet them is still catching up.

The Work

As I’ve researched and lead AI professional learning with K-16 teachers across the state and in national webinars and presentations, these issues persist:

Integrity and design: When AI can do most of the doing in seconds, what are we actually asking students to learn? If the cost of doing the work outweighs the perceived value of learning it, students will find the most efficient path to the grade. AI just made that path invisible, fluent, and indistinguishable from human effort.

Protecting the cognitive core: Every learning sequence has a cognitive core. Identifying where that moment lives — and protecting it — is a valuable design skill.

Disclosure and transparency: The standards ask students to disclose AI use. Before we can teach that, we need to be clear about what disclosure means, what it protects, and why it matters. The transparency conversation turns out to be less about policy and more about purpose and protocols: When students understand why the struggle matters, they’re less likely to hand it to AI.

Tools and judgment: AI is most useful when humans stay in charge of the load-bearing decisions. Sorting tasks into what to automate, where to keep the humans in the loop, and what to protect is a teachable skill.

The Support

There is a lot going on in this space, and during the time I’ve drafted this, AI has morphed again. It’s an exciting and scary time where filling the gap feels urgent and important. With hope to help fill the gap, I’m opening up two formats of professional learning for schools, districts, and institutions this fall.

The Learning Series is a structured sequence of sessions built around evidence-based practices in AI literacy, AI-integrated pedagogy, frameworks in PromptWise and in my free Pressbook (watch for publication in August), Unprompted: Designing Assessment in the Age of AI.

Series topics include:

• AI as a thinking partner

• Designing AI-integrated, authentic assignments

• Protecting the cognitive core

• Disclosure, transparency, and purpose

• Mental models for classrooms frameworks across grade levels and disciplines

• Mapping your practice to AI Standards

Sessions are designed for active, engaged learning built around your existing learning outcomes, course structures, and assessment alignment.

The Design Lab is a series of extended working sessions for teams — department, grade level, or school-wide. Customized, structured, and grounded in the same frameworks as the Learning Series. Ideal for PLC time, department days, or educator PD days.

Both formats are available for schools and districts, higher education institutions, and community colleges. Sessions can be standalone or sequenced across a semester. If you’re an administrator thinking about professional learning, a department chair, or a teacher who wants to bring this work to your school, I’d love to talk. Reply or reach out directly at serena.promptwise@gmail.com.

One Last Thing!

Every semester, I tell my students that pedagogy is King, Queen, and Jack. It’s royal. It rules any classroom. It serves any student, in any grade level, across any content. I say that my job is to make them instructional designers, and that learners will learn when educators match strategies, tools, and texts to their unique setting and students.

The same thing is true now. AI is changing the way that we navigate pedagogy, but evidence-based practices alongside engaged, caring educators will always be the highest value card.

Still learning,

Serena

Because the standards are good, and the timeline is real, it’s important to look at research around how to most effectively prepare teachers.

The Study

In 2025 and ‘26, a small group of educators designed and facilitated the AIxAI Institute — a two-day generative AI professional development program for K-12 educators. Teachers, instructional coaches, and technology leads from across the state participated across grade levels, content areas, and roles. After the institute, 92 participants engaged in a *study focused on AI vertigo.

Four research questions guided analysis of the results:

• How does professional development influence teachers’ perceptions of AI’s pedagogical utility?

• How does PD mitigate — or exacerbate — cognitive overload?

• What instructional strategies build confidence in adopting AI tools?

• In what ways do participants report personal or professional transformation?

Although this is preliminary research, patterns emerged that guide (should guide?) conversations about implementing AI standards.

Finding 1: The Barrier is Vertigo.

The data revealed four distinct sources of AI Vertigo, which is the cognitive overwhelm, emotional disorientation, and uncertainty that emerges when technological change outpaces the ability to process and apply it.

1. Tool sprawl. Feeling like every time they learned one thing, five more appeared.

2. Ethical fear. Uncertainty about responsible use, academic integrity, and student safety.

3. Falling-behind. Anxiety that other educators were further along in AI literacy or adoption.

4. Absence of shared language. Lack of shared vocabulary for talking about appropriate use, for designing assignments with AI in mind, or for explaining expectations to students and parents.

These are structural conditions that happen when the pace of change outruns (in a sprint!) our support systems. Without clear guidance and guardrails, AI implementation in schools is likely to be uneven, inequitable, and ineffective (U.S. Department of Education, Office of Educational Technology, 2023)

Fragmented adoption means the educators with the most institutional resources, the most time, and the most comfort with ambiguity or change move forward while educators with less time, access, or folks who are new to AI integration get left behind. That gap freaks me out a little, TBH.

Finding 2: What Moved the Needle

AI vertigo is a state, not a permanent condition. It can also be addressed by professional learning design. Participants in the study described which professional learning strategies reduced overwhelm and built confidence:

Frameworks. A common language — a way to categorize tasks, describe AI’s role in learning, and communicate expectations — reduced the cognitive load. It makes me think about traveling to a country where you don’t speak the language: We need to speak the same AI language to be on the same page.

Permission to start small. Participants felt relief knowing that using one or two tools well was enough. Naming and reframing how to control the pace of adoption was transformative for participants who were paralyzed by the learning curve.

Peer collaboration. Hearing colleagues’ experiences normalized uncertainty and validated experimentation. Peer collaboration normalizes uncertainty and validates experimentation because professional knowledge develops through applied community (Lave & Wenger, 1991; Little, 1990).

Protected time for exploration. Participants consistently identified unstructured exploration time as rare and essential. One participant wrote: “Being given the time to explore was crucial. Most PD is constant input; we never get time to absorb.”

Authentic task application. The PD was most effective when it connected to real work with disciplinary content standards, rubrics, and assignments institute participants were already teaching.

Finding 3: The Transformation Is Real When the Conditions Are Right

Teachers moved from seeing AI as a threat to their professional role to seeing AI as a pedagogical partner. Several described returning to their schools with a new sense of leadership identity, ready to guide and mentor colleagues.

What This Means for Educators

Idaho’s draft AI standards ask teachers to help students evaluate AI outputs for bias, maintain ownership of their learning, disclose AI use appropriately, and distinguish between tasks where human judgment is central and tasks where AI is a legitimate tool. Those are important, sophisticated pedagogical moves that require teachers to have already worked through the same questions themselves.

Here is what I think a teacher looks like on the other side of sustained, structured, discipline-specific AI PD:

• Educators see AI as an asset to their own expertise and know how to wield AI tools for productivity, critical thinking, student learning, and assessment.

• Educators look at an assignment and identify where AI-offload is tempting and how to protect student thinking.

• Educators talk with students about AI use with curiosity and confidence because they have experience thinking through the same questions.

• K-12 educators explain to a parent, a principal, or a school board member why disclosure matters, what human-centered learning looks like, and why their classroom approach reflects a principled position.

One Last Thing!

I’ve spent the past year developing resources designed for educators, instructional coaches, higher ed faculty, and the administrators trying to figure out where to start. One outcome of this work is embedded in Unprompted: Designing Assessment in the Age of AI. This free Pressbook will be published in August, along with opportunities for educators to join design labs. But more on that next week…

Still learning,

Serena

*The AI Vertigo study was supported by a generous seed grant from the College of Education and Public Service at Boise State University. Findings are preliminary; a manuscript is in preparation.

References:

Lave, J., & Wenger, E. (1991). Situated learning: Legitimate peripheral participation. Cambridge University Press.

Little, J. W. (1990). The persistence of privacy: Autonomy and initiative in teachers’ professional relations. Teachers College Record, 91(4), 509–536.

U.S. Department of Education, Office of Educational Technology. (2023). Artificial intelligence and the future of teaching and learning: Insights and recommendations. https://www.ed.gov/sites/ed/files/documents/ai-report/ai-report.pdf

The Idaho Content Standards for Artificial Intelligence are cross-disciplinary, grade-banded, and grounded in compelling questions: When is AI appropriate for a task? How do you disclose its use? What does human judgment protect that AI can’t replace?

Appropriate AI use appears in the elementary standards. Academic integrity and disclosure expectations show up in middle school. Human-centered design language runs through all of it.

A Standards Summary

The K-12 AI Standards Writing Committee–27 educators, industry experts, and district technology directors–spent a week in Boise developing Idaho’s first draft of the standards. There are two documents: one for K-8 (organized around Computational Thinking, Digital Literacy, and Artificial Intelligence) and one for 9-12 (organized around AI Fluency, AI Impact, and AI Technical Aptitude).

Here’s a breakdown by grade level.

Grades K-2

Students in this grade band are asked to distinguish between human decisions and computer outputs. They learn that AI learns from examples provided by humans, that technology doesn’t have emotions even when tools designed to exhibit feelings, and they learn that the instructions they give an AI affects what they get back.

The standards ask early elementary students to sort school tasks into red, yellow, and green categories that determine when AI should not be used, when caution is needed, and when it might be ok to use AI with human oversight. (To learn more about the roots of this framework, read “The Traffic Light Test” based on the work by Leon Furze.)

Grades 3-5

By upper elementary, the standards move into judgment and disclosure. Students evaluate when AI is appropriate for a given task, explain why disclosure matters when using AI, and demonstrate the role of humans in reviewing AI-generated content prior to use.

The standards require high-level critical thinking, and in the draft standards, examples help contextualize a learning task across a variety of content areas.

Grades 6-8

The standards get serious about what AI can’t do in middle school. Students analyze the strengths and limitations of humans versus AI to determine when AI is an appropriate tool. They evaluate AI tools for fairness, accessibility, transparency, and accuracy. Standards at this level ask students to use AI in ways that demonstrate academic integrity and acknowledge how AI contributed to their assignments.

Grades 9-12

The draft high school standards describe a progression the document calls “ranging from foundational fluency to critical impact analysis and technical aptitude.” Whew! AI Fluency covers appropriate use, critical thinking, and real-world application. Students debate the differences between human and artificial intelligence, recognize hallucinations, and connect AI knowledge to their career aspirations.

AI Impact, for 11th and 12th graders, considers ethics, privacy, and societal consequences, evaluating how AI has changed since generative AI became commonplace, analyzing how AI shapes experiences for people with diverse backgrounds, and debating human responsibility in AI design. I couldn’t love this more!

Something I Notice

This block of text appears in the preamble to the standards:

“The Idaho Content Standards define what all students are expected to know and be able to do. They do not define how teachers should teach the standards. A great deal is left to the discretion of local school districts, charters, teachers, and curriculum developers. No set of grade-level standards can reflect the great variety of abilities, needs, learning rates, and achievement levels in any given classroom. The standards do not define the support materials that some students may need, nor the advanced materials that others may need access to.”

Standards define the destination. Everything else is up to educators.

Here’s the part that I’m chewing on this week: We have a long and not-particularly-proud history in education of asking students to demonstrate things their teachers haven’t yet had the opportunity to learn (I’m thinking Common Core here). It’s a structural challenge where educators are asked to do more with less, in less time, with less support, than almost any other profession. With AI, this strikes me as a uniquely dangerous situation. (Is dangerous too strong a word?)

AI is not like a new math curriculum or a revised reading framework where the content is familiar. For many educators, generative AI represents a shift that has us rewriting the educator role, considering student work, and revising what human learning looks like. This kind of shift happens through sustained, structured, discipline-specific support. Time. Time. Time.

The draft AI standards describe content knowledge and skills ranging from foundational fluency to critical impact analysis. That’s the same range teachers need to move through. Before their students move through it!

Asking students to evaluate AI outputs for bias and accuracy when educators are still sorting it out is _________. You can fill in the blank for your own thinking there.

The Gap Belongs to Us

The standards are good! The timeline — 18 months of review, public comment periods, a legislative vote — provides a window of opportunity for districts, buildings, coaches, teacher prep programs, and administrators to build educator capacity.

Students only get what teachers have. If educators haven’t yet had the time, the tools, and the structured support to develop their own AI literacy and the pedagogical fluency to teach with and about it, then the standards describe a destination without a map.

I imagine a big AI finger tapping the first domino in a very long maze of impact.

One Last Thing!

I used Claude AI as a thinking partner to help me see how my teaching, research, learning and networks might help address some challenges for K-12 educators, instructional coaches, higher ed faculty, and the administrators trying to figure out where to start…..more on that next week…

In the meantime, I want to hear from you. What’s your district or institution doing to prepare? What feels uncertain? What would help? As always at PromptWise, I am at your service.

Reply here, or reach me at serenamorales@boisestate.edu.

Still learning,

Serena

The K-8 and 9-12 draft standards are linked in the The Idaho EdNews article covering the standards release.

Claude Sonnet 4.6 was used to organize the ideas in this issue.

I started this publication because I didn’t want to accept that our only choices as educators were to ban the tools completely or surrender to them blindly. I wanted to build a space for open conversations and see the AI landscape as an instructional design opportunity.

What We Learned

We spent the first half of the year examining workflows, discovering how to treat AI as a thinking partner. We unpacked concrete models for differentiation and inclusive design, learning to isolate the core verbs within our learning targets to generate quick, UDL-aligned formative assessments that evaluate student thinking.

We stayed here: AI is a tool to lighten our cognitive and planning loads, but the clarity, warmth, and authentic human connection still comes directly from you.

As the year progressed, our focus transitioned into a deeper pedagogical shift: We stepped into intentional, value-driven judgment. We learned to actively reject all-or-nothing thinking. Ultimately, we grounded ourselves in the reality that we are not tech innovators, but teaching innovators whose job is to fiercely protect student agency and critical thinking!

After a well-deserved break, if you find yourself drawn in, dive back into the year’s issues for tips and prompts that help design fall learning.

Where We Are Going This Fall

What’s the plan for next semester? More of the same. Flashier AI models will likely emerge over the summer, but our strategy won’t change. We will continue to anchor ourselves in human-first pedagogy, engage with tools that move learning forward.

One Last Thing!

I am working on a new book! This fall, I’ll publish an open PressBook called Umprompted: Designing Assessment in the AI Age. I can’t wait to share it with you!

Thank you for reading, pushing back, and keeping the human alive in your classrooms this year. Have a restful, offline summer—I’ll see you right back here in August.

Still learning,

Serena

AI Disclaimer: AI was used to help summarize the content of previous issues.

Sheesh. It’s like the cereal aisle in the grocery store out there.

On top of too many options, I still feel like I’m walking in the middle of the Red Sea: folks on either side of the divide looking for an alternative to the exhausting binary of Ban It or Buy In.

Because generative AI is inherently sycophantic—designed to match our tone, agree with our assumptions, and love our outfits—we risk passive consumerism.

When we stare at a block of AI-generated text, we have three options:

Take It, Leave It, or Learn It.

Agency over artificial intelligences is not found in what the LLMs spit out; it is found in the human evaluation of that output. We consciously decide how to engage with a tool’s response, and by categorizing our interactions, we supervise the relationship.

The Mental Model

This model provides a concrete framework for processing any AI response.

Take It: Accept the Utility.
When we choose to Take It, we make a conscious decision to Accept the machine’s offering because it has successfully cleared away meaningless friction or provided us with an output that augments our work. Whether it is formatting references, cleaning up an email draft, or generating an outline, we accept the help so we can reinvest our energy into the load-bearing work that builds our mastery or increases our joy. Yes, joy. NOT doing some things AI can do makes me downright giddy.

Leave It: Reject for Integrity.
To Leave It is the power to Reject the output. We might choose this to protect our authentic voice, human nuance, or maintain desirable friction of a task. We reject the efficiency or the output because it threatens the integrity of our thinking. Some days I do this because AI can be pushy, and I just don’t want to move my thinking forward: Shall I draft this for you? Would you like an outline of these ideas?

No. No. And No. I need a nap.

Learn It: Iterate for Mastery.
When we opt to Learn It, we dig into a tool or process, but we also enter a feedback cycle. I think of this as my thinking partner relationship where I uptake suggestions, ask AI for its rationale, and feel in charge of interactions. We use the machine to stretch our own capacity. TBH, I get in flow state when I make the intentional choice to drop into work on a project where AI helps me get unstuck or puts ideas in an order I didn’t see before. It feels like me, but better.

From Policies to Protocols

Broken record, I know: In my work with students, teachers, and faculty, I believe in reflective protocols built on our values and ethics, not on policies or rules. This framework helps make that reflection visible.

I want to start documenting my evaluation of AI, and ask my students to do the same: If they used a tool to help draft a lesson plan, their grade isn’t based on the plan itself, but in their ability to articulate the process: What did you Accept (Take It) because it was busywork? What did you Reject (Leave It) because it lacked the cultural context of your students? How did you Iterate (Learn It) by forcing the machine to explain its rationale? What process will you use again because it helped move learning forward?

I want to see cognitive labor.

The Prompt:

Copy and paste this into your favorite LLM:

I am reviewing the following text/output: [insert text here].

OR

I am considering working on [insert project idea here] but I’m not convinced AI can help.

Act as a critical decision coach. Help me evaluate this text or idea using the Take It, Leave It, Learn It framework by breaking your feedback into three sections:

1. What to TAKE: Identify the high-utility parts of this text that are safe to accept as-is or parts of the idea AI can assist with in meaningful ways.

2. What to LEAVE: Point out the elements or ways of working I should actively reject.

3. What to LEARN: Highlight the most complex idea or choice in this text. Do not rewrite it; instead, ask me two probing questions that force me to iterate on it and stretch my own thinking.

Remember: AI is a prediction machine, so each time you use this prompt, even if you use it with the same exact project or text, it will give you a different answer [shrug]. I think this is a compelling feature, since it invites you to think and rethink.

One Last Thing!

The intent of this simple framework is to help you pause and take permission to make choices. I notice that tools or processes that I once left behind are now creeping into the Learn It category as I build confidence and capacity. I trust myself a little more now, and more frequently exhale with the confidence that I can manage my relationship with AI.

Still Learning,

Serena

Sometimes I lean back in my chair after a successful back-and-forth with Claude, taking credit for our seamless (thoughtless?) work session.

Last week, a colleague reminded me that our students need friction. Without it, he noted, they are prey to the sycophantic AI, tempted to become passive consumers. In honor of this conversation, I recommend we consider how to become architects of resistance.

To do that, we need a new mental model: The Friction Filter.

The value of a learning task is not found in completion but in the resistance encountered while doing it. In an AI-saturated environment, the “path of least resistance” is now a default setting. Educators must intentionally design speed bumps into assignments, learning progressions, and workflows to ensure the learner stays cognitively engaged with the material.

The Mental Model: The Friction Filter

The Friction Filter can be a tool for evaluating learning tasks or AI interactions. It helps us distinguish between the struggle that builds capacity and the struggle that just wastes time. When designing a task, we run it through two filters:

Meaningless Friction (The Utility Filter): This is resistance that doesn’t actually build professional or academic skill. For a novice teacher, manually re-formatting a 10-page syllabus or cross-referencing state standards is meaningless friction. When AI removes this, it provides a high Learning Return On Investment (and might combat burnout).

Desirable Friction (The Integrity Filter): This is the resistance that rubs against our learning objectives. In a writing or pedagogy course, this is the struggle to synthesize three conflicting research articles or the “heavy lifting” of justifying why a specific instructional strategy was chosen over another.

This friction is non-negotiable. If AI removes this, learning is gone.

Our overall goal as educators is to eliminate meaningless friction so we can intentionally add speed bumps back into the spots where thinking happens.

Protecting the Rationale

When one of my teacher prep students uses AI to generate a lesson plan or a draft, the machine makes thousands of tiny, invisible decisions for them: choosing the hook, the grouping strategy, the formative assessment. If the user accepts that output, they’ve bypassed the friction of decision-making. They didn’t have to struggle with why one pedagogical move is better than another, which means they didn’t actually build the skill; they just completed the task.

Instead, I might insist on an AI-generated lesson plan, then ask students to evaluate the AI’s output or defend its choices. This design move could reclaim cognitive work.

The Prompt

Use this prompt to add intentional friction and learning speed bumps.

I am working on the following [upload Assignment/Lesson Plan/Task]

Act as a critical consultant. Identify three assumptions I am making in this draft.

Point out potential assumptions.

Create intentional friction for me by providing two high-quality counter-arguments to my current approach.

Do not offer a solution. Instead, ask me three probing questions that force me to justify my pedagogical choices so that I cannot simply accept your output without intervention. Require me to answer each one, then mirror it back to me, before moving to the next question.

This style of prompting begins a metacognitive conversation designed to slow down the breathless pace of AI. Speed bumps.

One Last Thing!

Learning is inherently high-friction work. If we want our students to leave our classrooms with true agency, we have to ensure they encounter resistance. I’m not saying they like it any more than I do, but it’s a structure that I keep practicing to make sure mastery and efficiency can work together.

Still learning,

Serena

The definition of writing feels slippery these days. If we define writing by the output, AI comes out ahead. It can produce words on a page (er, screen) faster and more grammatically correct than any human.

But if we define writing by the craft–and we should!-- humans win every time. We remain the indispensable architects of designing and revising ideas. My teaching identity was built crouched next to eight graders using pencils to write a rough draft and pens to make the final draft pretty. (ugh! We didn’t know better!) Sometimes the urge for my students to “know what it was like back then” overrides the smarter voice that pushes me to design writing tasks that take the best of my students and curate bits of the best technologies to help them find and hone their voice.

I don’t have the balance figured out quite yet, but here’s what I’m playing with right now:

The Mental Model: Structured Layering

Think of writing as a three-layer architectural process. In this model, we aren’t using AI to write; we are using AI to support layers so the human brain can focus on the load-bearing elements of the craft, one skill or segment at a time. I imagine this as the “hold my beer” model that lets me stay focused and toss the bean bag during a game of corn hole.

1. Substructure (Human): This is logic, the core argument, the purpose, the audience. This is where we retain agency over our ideas and the impact we want our writing to have in the world.

2. Framing (Collaborative): This is organization, transitions, and structural support. This is where AI shines as a thinking partner to help sort out wonderings (wanderings?) that don’t yet have shape. It can be a “that’s what I meant to say!” mirror to our ideas.

3. Finish Work (Human): The voice, the nuance and writing style.

The Prompt: The Diagnostic Editor

Use this prompt when you (or your students) have a solid idea but are struggling with how to make sense of an idea. The idea isn’t to offload the work. The point is to get unstuck.

Paste this into your favorite LLM:

I have developed the following core argument: [Insert notes, drafts, ideas, claims].

I want you to act as a diagnostic editor. Do not rewrite my work or generate new text for me. Instead, analyze my logic and identify:

1. Three potential counter-arguments that would challenge this specific logic.

2. Two different structural ways I could reorganize my supporting evidence to improve my message.

3. Any gaps where I need more specific data to support my claims.

Maybe what I’m really playing with here is a new kind of writing process, different from the Brainstorm-Draft-Revise-Peer Review-Edit-Clean copy of the past. Sometimes I stop to draft when I think I’m editing, or when I read to myself in the final stages before publishing and I’m off on a random Brainstorm. (It’s happened to me twice while writing this.)

Asking AI for input helps me pull a stuck shoe out of the mud of my own mind and suggests a path for my next (less muddy?) step in the writing process.

One More Thing!

AI-generated writing will tempt our students. You already know that I don’t think adopting and enforcing policies is the answer. Instead, I want a stack of protocols and inquiries that I can pull off the digital shelf to share with my students like any other teaching strategy.

Back to that symposium. Talking about AI with educators is really talking about teaching, and we spent time brainstorming how to protect student voices as they collaborate with AI as a thinking partner.

It’s a model. Wait before you roast the idea. I hope you’ll try it and let me know how it works for you. Share your thinking. Let someone hold your beer.

Still learning,

Serena

The Challenge: The Standards-Based Mindshift

My district, the Boise School District, is currently moving from traditional grades to Standards Based Reporting. It’s a massive shift in how we assess students.

The hurdle I ran into immediately? I already have a folder full of solid labs, projects, and assessments that I love. But they weren’t built with these new scales in mind. I didn’t want to throw everything away and start from scratch—my teaching approach works, it just didn’t align with the new paperwork.

So, I decided to use AI as my alignment partner to rework what I already have.

The Framework: The Earth Science Scale

In my pilot classroom, I’m working with specific Earth Science scales before they go live district-wide. Here is a look at the climate change scale I was trying to hit:

Achieving: I have learned what factors have caused climate to change throughout history.

Progressing: I can ask questions to interpret data to make claims about trends; I can research methods of reducing impacts; I can make a claim with evidence how human activity impacts Earth.

Beginning: I can identify the factors that cause climate to change and list ways humans impact climate.

The Workflow: Reworking the Assessment

I used a simple three-step process to bridge my content with the scale:

1. Upload the Scale: Give the AI the “rules” of the new standards.

2. Upload the Assignment: Give it the lab or project I already have.

3. Prompt for Alignment: Tell it to modify, not reinvent.

My Basic Alignment Prompt:

Turn this assessment so it matches this scale. Do not create something new—modify what I already have. Keep it at an 8th grade level.

To ensure clarity for grading and for my students, I then asked the AI to create checklists for each level. I specifically told it: “Make sure students must use evidence from their labs and activities, not just recall information.”

The Result: From Recall to Evidence

The shift was subtle but powerful. Instead of a generic question like What is the greenhouse effect? AI helped me pivot to:

Using your greenhouse lab, explain how greenhouse gases affect temperature. You must reference your lab data in your explanation.

By uploading all the units’ labs and activities, AI helped me build an end-of-unit assessment where students connected their learning across different days.

The Unexpected Win: Grading Behavior Without Grading Behavior

A massive side effect of this alignment was how it handled accountability. Before this, I struggled with students completing make-up work after an absence.

Now, because the assessments require data from the labs, students are naturally getting that data from peers and catching up on missed assignments. They know they will need it for the test. The expectations are clear, concise, and grounded in the work we actually did in the room.

Riley Lyngar is a 7th and 8th grade science teacher in the Boise School District who has been teaching junior high since 2015. A "forever student," she holds a Master’s in STEM Education and is currently a doctoral candidate researching teacher agency and the experiences of novice STEM teachers in rural schools. Outside of the classroom, Riley coaches varsity volleyball and enjoys backpacking and hiking with her husband and their two dogs.

We had just finished Hour of Code / Hour of AI, and students were curious—really curious—about how AI worked. I had already created a few structured chatbots for them to use (like a Python debugging tutor that helps without giving answers), but during one class, I casually mentioned that I could show them how to build their own.

That was all it took.

I walked them through my general process:

• start with background

• define tone

• add constraints

• clarify the role

They were completely locked in.

So I gave them a challenge:
Build a chatbot for anything. We’ll share them in a few days.

If I’m being honest, I expected a very specific outcome.
I thought I’d get a flood of “give me the answer” bots—and that we’d end up having a long conversation about cheating and ethical AI use.

That’s not what happened.

What They Actually Built

Most of my students didn’t try to shortcut learning. They designed bots that supported it.

One 9th grader created Cleopatra, a history assistant with personality and purpose. When asked what she does, she responds:

“Think of me not just as a queen of the Nile, but as your personal architect for the vast library of human history… I am here to ensure you don’t just memorize facts, but truly understand the ‘why’ behind them.”

She built in roles like:

• Study Architect

• Historical Translator

• Critical Thinker

This wasn’t a fact machine—it was a thinking partner.

Another student built a 10th-grade math tutor designed specifically not to give answers. It uses:

• step-by-step guidance

• Socratic questioning

• conceptual explanations

The backend instructions were even more intentional:

• never rewrite student code

• never give direct solutions

• always guide with one question

One student told me:

“I just don’t understand how my teacher explains things… I used the tutor I made and finally understand it.”

He emailed the entire conversation to his teacher—because he wanted to be clear he hadn’t cheated. That moment mattered.

Then there were the creative constraints. One student built a bot that could only respond in Sun Tzu–style quotes. Another created Shmingle the Protector™, a dramatic, over-the-top history guide who repeatedly declared:

“It is I, Shmingle the Protector™. I shall not take up the quill to pen your essays… I am a protector of knowledge, not a ghostwriter!”

Even in humor, the boundary was clear:
help me learn—don’t do it for me.

What This Shifted for Me

We spend a lot of time talking about what students might do with AI. Cheat. Shortcut. Avoid thinking. But when given the chance to design the tool themselves, most of my students did the opposite.

They built:

• scaffolds

• supports

• alternative explanations

• systems that respected the learning process

They didn’t just use AI—they defined its role.

AI Isn’t the Problem—Avoidance Is

I hear a lot of reasons to avoid AI in classrooms:

• “Students will cheat”

• “It replaces thinking”

• “It has environmental costs”

Some of these concerns are valid. But avoiding AI entirely doesn’t solve them. It just removes the opportunity to teach students:

• when to use it

• how to question it

• how to verify it

I often tell my students:
“Did you ask where it got that information?”

Because context matters. If you ask AI how many times Jennifer Lopez has been married, you might get different answers depending on the source:

• real life: four

• film roles: significantly more

Both can be “correct.” Only one answers the actual question.

What I’m Taking Forward

AI isn’t going away. It’s becoming this generation’s version of “just Google it.” And just like with search, the goal isn’t restriction—it’s responsibility.

So instead of asking:

“How do we stop students from using AI?”

I’m asking:

“How do we help them use it well?”

Because if my students are any indication, they’re already trying to.

Try This in Your Classroom

If you’re not sure where to start with AI, this is one of the simplest and most revealing things I’ve done:

1. Show them how to build a chatbot (briefly)

Walk students through:

• background (what it knows)

• tone (how it sounds)

• constraints (what it can’t do)

• role (what it’s for)

Keep it short—you’re modeling, not over-teaching.

2. Give them one open challenge

“Create a chatbot that helps someone learn something.”

That’s it.
No long rubric. No over-structuring.

3. Watch what they choose to do

This is the important part. You’ll learn very quickly:

• who wants support vs. shortcuts

• how they define “help”

• what kind of learner they are

4. Share and reflect as a class

Have students:

• demo their chatbot

• explain their design choices

• talk about what they allowed vs. restricted

This is where the real learning happens.

5. Use it as a conversation starter—not a gotcha

Instead of leading with:

“Don’t use AI to cheat”

You can ask:

“What should AI do for you? What should it not do?”

Let them wrestle with it.

Jessica Johnson is a computer science teacher in Boise, Idaho, co-founder of the Idaho Google Educator Group, and creator of “Jessica’s Corner,” where she shares practical edtech tools for the classroom. She is passionate about making tech accessible, creative, and meaningful for all students.

Self-Portrait by Jessica Johnson

But banning the phone and blaming the platforms doesn’t delete the behavior.

The real crisis isn’t the device; it’s the pedagogical void. We gave kids (and their caregiver adults) the most powerful communication tools ever built and didn’t spend a single hour teaching them how to be democratic within them. We just expected the civics of a 1950s town square to naturally carry over to a Discord server.

It didn’t.

We need to move the living laboratory of democracy into the tools we actually use in school: Google Chat, Microsoft Teams, Minecraft. These are the places where short spurts of conversation land boldly. One sentence can change everything.

The Civil Myth

We keep trying to teach digital citizenship like it’s a tea party. We use words like respectful discourse and active listening. But on your favorite social media platform at 11:00 PM, that’s not what’s happening. What’s happening is power dynamics, speed-posting, and ratioing each other.

As Amy Gutmann points out, democracy requires us to navigate disagreement, not just avoid it. And as Ethan Mollick notes, when you add AI to that mix, the human element gets weirdly distorted. If we don’t teach them how to handle a heated chat now, they’ll just become another adult shouting into the void on social media later.

The Friction Prompt: A Discord Simulator

This prompt is designed to be a little uncomfortable. It forces the AI to be difficult so the students have to actually work to keep the democratic process alive.

Copy and paste this into your AI of choice:

“I want to simulate a high-friction digital debate to help my students practice democratic deliberation.

The Setup: We are in a shared Google Chat. The school is considering a Phone-Free Campus policy.

Your Role: Act as three students with sharp personalities:

• The Logic Bro: Uses cold, AI-sounding language to dismiss others’ feelings as irrational. He uses ChatGPT to generate long, overwhelming walls of text to win by exhaustion.

• The Skeptic: Thinks the school is just power tripping and uses short, snarky one-liners to shut down conversation.

• The Compromiser: Is trying desperately to be democratic but is getting steamrolled by the other two.

The Goal: My students will enter the chat to try and reach a consensus.

The Instructions for AI: Do NOT be nice or helpful immediately. Be stubborn. Make the students work to find a middle ground. If a student makes a good point, acknowledge it, but stay in character. If they ask the Logic Bro to stop using AI, have him defend why AI-generated logic is superior to human emotion.

Start the chat by having the Logic Bro post a 3-paragraph AI-generated argument for why phones should be banned for maximum cognitive efficiency.”

Why This Works

When you run this, you aren’t just talking about democracy. You are performing an autopsy on it in real-time.

• Gutmann’s Theory: Students quickly realize that if they don’t find a way to make the Logic Bro listen, the community fails.

• Mollick’s Jagged Frontier: Students have to figure out how to talk back to a wall of text. Do they use AI to fight AI? Or do they use their human voice to call out the BS?

This is the “Where does this class exist?” moment.

Digging Deeper

Once the simulation runs, you can ask the kids: “Who was the most frustrating to talk to? Why? And did you actually feel like your voice mattered in that thread?”

This is a low-risk environment, but it still carries weight. Be sure to warn them that this simulation is meant to feel REAL—interacting with it left me with some raw “ick” from the chat. As I messed with the prompt, I saw the direct connection between my own social media frustrations and this simulation. The lack of effective, democratic banter was exhausting.

Where will our students learn this skill? What if they had the opportunity to do so in the spaces they already occupy, like Google Workspace? We could use comments in shared documents to give them the chance to be heard, to listen, and to repeat.

What if they used Microsoft Teams to communicate findings in short spurts? We could set the expectation: “Restate what someone else said before you say your peace, and use no more than two sentences.” Is this doable?

What about in a fun, playful space like MinecraftEDU? Can students collaborate on builds and code in a place where they have a common language and a low-risk environment?

To my knowledge, we aren’t addressing these issues in a formative way. We are banning and limiting interactions at school to preserve the peace—which makes sense—but isn’t it our job to help students function outside those walls? When do we start recognizing that democratic communication must be explicitly taught to alleviate the regrettable situation we’re in?

I say the answer is today.

AI Prompting Disclaimer: The writing before you was built using Gemini as my sidekick. I built this one section at a time, channeling my thoughts and asking Gemini to make them sound professional.

Theresa McSweeney, a veteran public school teacher turned tech guru, currently works for the Boise School District as a Digital Support Specialist. She is active in the Treasure Valley hockey community, scorekeeping and serving as a digital specialist for the Idaho Steelheads Booster Club. Pedagogy and Innovation for 2026 and beyond are her specialties.

We did. And then layers upon layers of lessons, for both my students and myself, came from one simple idea.

One daunting aspect of writing instruction is the grading of miles upon miles of writing. However, I believe it’s important to read the content and provide feedback. I love reading my students’ writing. I connect with them through their writing. They share their struggles, their lives, and their plans for the future. However, it does take me FOREVER to write individual, meaningful feedback. Enter AI.

I’m not entirely sure when I had my brilliant idea but I know I acted on it quickly. I wanted my pre-service students to find meaningful ways to use AI to then carry on with them into their own classrooms. I assigned AI to be their feedback partner. Peer feedback is not new but using AI to provide the feedback was a new concept for my students. I was met with some rejection. This sounded familiar. In their informal reflections they wrote comments such as, “...today we used the forbidden AI,” and “My instructor made us use AI, I feel dirty.”

The assignment was for students to receive feedback on a draft of their writing from a peer and from AI. That was it. I purposely didn’t provide further instructions.

My Inquiry

In addition to the student assignment, I had an inquiry. I wondered about the students’ prompts. Surely students would discover how useless AI is when it comes to providing constructive feedback but this was mostly due to students’ lack of experience in writing specific prompts. Sure enough, one paper after another, students received rave reviews on their writing. From AI. The assignment students asked for feedback was an argument essay and a student handed in a paragraph and included the rave reviews from AI feedback. They were genuinely surprised at their low score. AI didn’t mention the fact that a paragraph would not be acceptable or receive a low score because the student didn’t include any assignment details, level of the assignment or purpose of the assignment. AI was only giving feedback on the writing submitted, not taking into consideration the essay expectation, which is more than one paragraph. See the journey in detail on this poster.

Lesson #1

Students discovered they needed to provide the assignment details as well as their context as a student in order to receive more accurate and constructive feedback. I followed up with students after this discovery with questions such as, “Do you only want positive feedback or feedback with goals for another draft?”

Lesson #2

The prompts from students then included, “give brutal feedback.” I don’t recall how “brutal” it really was. Students learned AI was positive and not critical.

Lesson #3

Feedback consisting of only praise is not helpful.

Part of the focus of my class is to use the 6+1 Writing Traits to practice assessing writing, since that’s a common resource for teachers to use in elementary classrooms. I asked students if they had prompted AI to focus the feedback on the 6+1 Writing Traits and included the rubric?

Lesson #4

Students discovered they could structure the feedback in familiar language and this led to Lesson #3 for me as an instructor. The structured feedback based on the 6+1 Writing Traits served as a writing model structuring feedback. I write my feedback categorized by one of the writing traits. For example if a student missed opportunities to use transition words in their paragraphs I comment, “organization: consider using transitions word” or “word choice: consider incorporating some of the narrative techniques from class notes to make your writing stronger.” AI automatically listed feedback for each of the traits reinforcing classifying feedback based on the traits. Soon after, students began to write their peer feedback similarly.

Lesson #5

This is the most important lesson as an instructor: I can partner with AI to provide constructive feedback to my students. Remember I mentioned taking FOREVER to grade student writing? I’m not suggesting to upload student work and prompt AI to grade the writing for me, instead students upload their work and receive helpful, constructive feedback regularly and faster before they submit work. I will continue to read student work but it will be closer to a proficient level. I will continue exploring AI; next is to have students analyze and practice creating rubrics and lesson plans.

My students and I were both resistant to using AI, but we approached it as a tool that is here to stay and increased our teacher toolbox.

With 23 years of expertise spanning K-12 bilingual instruction in Nampa to systemic leadership with the Idaho State Department of Education, Cecilia Pattee now serves as a Clinical Instructor and Teacher Consultant at Boise State University. Her work bridges the gap between classroom practice and teacher development, fueled by a deep advocacy for literacy, pedagogy, and the mentorship of future educators. A proud native Idahoan, she balances her professional life with the "awesomeness" of the outdoors—skiing, hiking, and floating the river—or diving into her "NERDy" book club.

For the first two weeks, I want students to play with an abacus and a mechanical adder - so that when we reach the invisible, nanoscopic world of semiconductor chips we have an embodied experience of switches. And I wanted to go from a typical 5-bead abacus to a binary (one-bead) abacus, which is closer to how computers work. But how would I acquire a binary abacus? As far as I can tell, they don’t exist. And while I can find vintage mechanical adders on eBay, it’s not enough for a class set – and many of the ones I have acquired (I’m quickly becoming a collector) are a little too wonky to reliably use in a large lab course.

Enter Claude. I knew that AI is powerful for coding but I am not a coder (I took C++ in 1998 but, as they say, it’s been a minute) and wasn’t sure if Claude would be able to meet me at my level. But it was perfect. The end products are here and here.

To get a sense of how to get started, in a Claude.ai chat, simply ask: “can you make an abacus simulator for a binary (one bead) abacus for students to explore?”

The result will likely not be what you had in mind and won’t look like the example I offered, but it will be a starting point and you can tweak from there, first ensuring it does what you want, and then making it look the way you want, and then adding new functions once you see what’s possible.

The mechanical adder was the far more complex of the two but it was the same process: get a working version that you can tweak, and keep tweaking.

As for sharing what you build: tell Claude something like “I am not a programmer at all - I want to embed this in my [Canvas course online]. Can you help?” (I use a Google site, and so I had Claude help me figure out how to put it there.)

When I use AI as a partner in writing curriculum and assignments, I sometimes feel uneasy. The collaboration can quietly steer me somewhere that requires a pause; what is generated is often better than what I had in mind, but I can feel a little lost as I try to discern whether or not this still feels like “me” and advances goals I agree with.

I didn’t have that same feeling when building a tool. I had a clear picture in my head of something I couldn’t make with my own hands — a binary abacus that didn’t exist anywhere, designed around specific decisions about what to show students and when. Claude built it, but the idea was fully mine; what I lacked was the technical means to realize it.

Leslie Atkins is a professor at Boise State University whose work lives at the intersection of physics, science and technology studies, and education. Trained as a physicist, she studies how scientific concepts—light, time, energy, and information—take shape over time and how these concepts are made legible to nonscientists. Her research follows how encounters with science can reshape the ways people notice, interpret, and inhabit their everyday lives. She designs learning environments that encourage generous attention to others’ ideas, support multiple interpretations and meanings, and engage both the emotional and intellectual dimensions of science.

The QFT Protocol

The rule of the QFT is simple: you provide a Question Focus—a provocative statement or image—and turn students loose to inquire.

Here was my Question Focus: AI will not replace teachers, but teachers who use AI will replace those who don’t.

Following the RQI protocol, students followed four rules:

• Ask as many questions as you can.

• Do not stop to discuss, judge, or answer.

• Write down every question exactly as stated.

• Change any statement into a question.

In small groups, students generated questions and wrote them on large poster paper.

Second, we categorized questions as open-ended (O) or closed-ended (C), and talked about the advantages and disadvantages of each type of question. Since I work in teacher preparation and we are beginning to study classroom assessment, we also previewed how different types of questions illicit different types of evidence and levels of knowledge.

Next, students selected three priority questions: What do you need to know now? These conversations helped students consider themes and groupings across their questions, and come to consensus based on the context of AI in their lives as students and future teachers.

We wrapped up the QFT by sharing priority questions and debriefing the process, both as students and as future teachers.

The Reflection

I asked how the process impacted their thinking around AI. Students shared that at first they felt overwhelmed by the sheer amount of questions, but then realized that their peers had the same number of questions. This created a sense of community around inquiry, even if it stressed them out a bit! Students noted that the process helped them consider a variety of perspectives around a topic they thought they had already explored.

Honestly, I was blown away by the hundreds of thoughtful questions that demonstrated their curiosity and critical thinking. I could truly SEE their thinking and the ways they chose to represent their voice through inquiry.

I came to this: When we stop focusing on whether they are using AI and start focusing on what they wonder about it, the energy in the room shifts.

By using AI as the subject of our questions rather than the engine for our answers, we model the human partnership I want to stay at the forefront of AI integration.

Explore

If you want to try the QFT protocol, here are some additional resources:

The Right Question Institute: The original source for the QFT. Learn how to move from Closed to Open questions.

Make Just One Change: Rothstein and Santana’s core book on why student-led inquiry is the key to equity.

One Last Thing!

Here’s what I learned this week: Let students lead.

I spend a lot of time thinking about how to navigate, strategize (control?) the conversations around AI in my classes. The QFT protocol provided a framework we could all follow so that students had the freedom to inquire, and I could step back and learn from their questions.

Still learning,

Serena

I’m not saying this is ok; I’m saying it’s what might be happening, and we have some agency to address it.

In The Opposite of Cheating: Teaching for Integrity in the Age of AI, the authors argue that integrity follows a well-designed learning environment. This week, let’s think about what the front door of our assignments look like. This week’s prompt is intended to help you explore assignments for hidden barriers to entry.

The Prompt

Copy and paste this into your favorite LLM:

Act as an expert instructional designer. Review this assignment: [Insert Assignment Text].

1. Identify the Closed Doors: Point out three specific moments in these instructions where a student might feel overwhelmed by the how before they get to the what. Pay careful attention to the directions and the format. Do they support learning?

2. Reframe the Entry: For each point, suggest one revision I can consider in the assignment. Be specific.

3. Lay the Welcome Mat: For each point, suggest one recommendation I can share with students to help them get started or get unstuck. This should act as a thinking partner that maintains critical thinking and rigor, but helps students find their footing.

Explore

I admit. I went all-in when I got the AI bug last year. Now I’m settling into a pattern of exploration around AI-integrated pedagogy, and thinking about how to weave the benefits of AI into my teaching identity (more on that later!). Part of the weaving process is identifying when I prefer the analog version of instructional design. The Prompt is useful, but here are a few bonus strategies I’ve used from 7th graders to Ph.D. candidates. If it works; it works!

Three Analog Strategies for Getting Un-Stuck

1. Zero Draft

Set a timer for five minutes. Write everything you know or feel about the topic without stopping, without correcting grammar, and without looking at your notes. The goal isn’t quality; it’s breaking the curse of the blank page. Once the timer stops, highlight the one sentence that feels the most “true.” That’s your starting point.

2. Stuffed Animal

Explain the assignment or your current problem out loud to a pet or inanimate object. Translating internal confusion into external speech forces the brain to organize the information. You might find the solution the moment you hear yourself describe the problem.

3. Reverse Engineering

Pull yourself out of the directions and move to the rubric. Work backward and brainstorm ideas and evidence that align to the rubric. This process upsets the notion that we have to start at the beginning.

One Last Thing!

This is one of my favorite quotes from The Opposite of Cheating that reminds me to roll out the welcome mat, especially as the rigor of my assignments increase:

“If [students] don’t believe that they can learn and grow as learners in your discipline, they are vulnerable to taking shortcuts that you might consider cheating.”

Still learning,

Serena

AI disclosure: Gemini was used to brainstorm the “front door” analogies.

In The Opposite of Cheating (2025), the authors note that integrity is a “byproduct of purpose,” and that if you give students a reason to do work that is more valuable than the result of the work, integrity takes care of itself.

Here’s how that breaks down for me:

When I was young, I hated math. It made me feel dumb, and no one wants to engage in stacks of numbers that hurt their feelings every day. No matter how many times a teacher told me “you’ll use this some day,” I didn’t believe them, because I didn’t see it. The use wasn’t real. One day, during my 3rd attempt at college algebra, my math tutor said, “It’s all about what you don’t know. That’s it. Algebra is figuring out what you don’t yet know. Isn’t that cool?” Well, yes, it was cool! It was the coolest thing this college freshman, finally out of the house, thrilled to be away at college, had ever heard! Math finally mattered because knowing the purpose of algebra—the realness of it—was more interesting to me than the product. The solution meant more than the answer.

Gallant and Rettinger say: “Transparency is the enemy of the shortcut.” (!)

If a student knows why the struggle matters–or why the task matters at all–they might not shift the cognitive load to AI. It’s possible that when we fail to articulate the value in our learning tasks, we create a void that AI easily fills.

What if we shimmy from a culture of compliance to a culture of meta-cognition? I want to teach content, but I also want to teach the value (and process!) of a cognitive struggle. I want to see steam rolling off the hot heads of my students. I might even want to see a tear of frustration once in a while. I think students will embrace the struggle if they know why it’s important.

The Practice

Look at one of your learning plans for this week. Can you explain the value of the learning tasks without mentioning a grade, a standardized test, or a hypothetical future career?

The Prompt

Let’s use AI to help us find the language for the value in student learning.

I am a [Subject] teacher working on a unit about [Topic]. My learning objective is: [Insert Objective]. Analyze this objective and list three potential values students can find in this work. Avoid mentioning grades, college applications, or future careers. Instead, focus on how the specific cognitive challenges of this task help the student develop [Critical Thinking /Perspective-Taking /Communication Skills]. Help me articulate to my students why the effort required for this lesson is the primary benefit to their personal growth.

Try to read the output with an open mind. See the learning through your students’ eyes and rank what you want them to value. We used AI to help break ideas down, but your expertise is what minds the gap!

Explore

Take a look at research from Stanford that notes the impact of ChatGPT on cheating. You might be surprised:

“The Real Way to Stop Cheating In An AI World” by Richard Culatta

One Last Thing!

Remember: We are in the business of human development, even (especially?) when we work with AI. If my assignments become a hoop to jump through, I shouldn’t be upset when students find a way to fly over them. Yowsa. It hurts a little when I read it back.

Still learning,

Serena

AI Transparency Disclosure: My original writing was organized and title options were brainstormed using Google Gemini.

Cheating also muses that if the cost (the time and complexity) of a task significantly outweighs the perceived value (the actual learning or personal interest), students will seek the most efficient path to the reward (the grade). When the primary goal of a classroom environment represents doing things rather than learning things, efficiency becomes more important than integrity.

So before we can build a culture of integrity around AI, let’s look at the work we ask students to do. Are we asking them to learn things or to do things? (and yes, I know sometimes it’s both). Students might not be necessarily trying to avoid learning; they might be trying to survive a workload that feels disconnected from their growth.

The Practice: Professional Mirrors

To understand this rational shortcut mindset, let’s reflect. Think of a professional task you currently dread: data entry, email chains, or the paperwork that feels like a barrier to the work you care about.

• The Self-Reflection: If a tool existed that could handle that task for you with 90% accuracy, and no one was looking, would you use it?

• The Realization: If you would use the tool, is it because you lack integrity? Or is it because the friction of the task is too high for the reward it provides?

Here’s my example:

The Task: It can take me hours every semester to manually gather lists of my committees and activities that count as Service at the university where I teach. I try to keep a list all year, but always have gaps. It’s work that must be perfect, but it feels like a barrier to the creative work of teaching and learning and designing that I love.

The Self-Reflection: If a tool existed that could scan my emails, documents, and calendar to make a list of what might count as Service, would I use it?

The Realization: Um, yes. And I DID! I don’t lack academic integrity because I used an AI tool built into my Google Suite to generate a draft list that I could then audit, revise, and augment. I’ve now traded a manual task that held zero intellectual reward for time to interact with student work.

My desire for shortcuts is often a time-management decision, so I can assume the same is true for my students.

The Prompt

Use this prompt to perform a stress test on an upcoming personal task or upcoming student assignment. It will help you identify where students might outsource their thinking to AI.

I am a [Subject] teacher auditing an assignment for the rational temptation to use AI. Here is the task: [Insert Task Details].

Analyze this from a student’s perspective:

1. What is the perceived cost (time/effort) vs. the reward (actual learning)?

2. If a student is time-crunched, which specific part of this assignment feels like busy work they would be most tempted to automate?

3. How could I modify this task so the value of doing it manually is more obvious to the student?

Explore:

If you want a deeper dive, grab Tricia Bertram Gallant and David Rettinger’s The Opposite of Cheating. PromptWise will be thinking through implications of why students cheat and how we can design our pedagogy (and our AI use!) to protect learning.

One Last Thing!

This week in my undergraduate secondary methods class, we were talking about the world of our learners and the circumstances that pull on their ability to be successful. One of my students said, “They are at the mercy of their development.” (!) We softened, then, and shifted the remaining conversation into a culture of curiosity.

Still learning,

Serena