
Beyond the Stoplight: Rethinking AI Use Through a Lens of Student Agency in Mendon-Upton
A Case Study from the 2026 PEA2K cohort
Most schools treat academic integrity as a disciplinary conversation. There are rules and consequences and the conversation centers on whether a line was crossed.
Ryan Robidoux, Director of Learning and Innovation at Mendon-Upton Regional School District, believes that there is a better conversation to center as districts build frameworks for student AI use. "Acceptable use policies tell students what to do," he says. "They hand down rules. They check boxes. And in doing so, they quietly remove the very thing education should be building: student agency."
Mendon-Upton Regional is a small regional district serving two towns in central Massachusetts. The district's instructional framework centers deeper learning, with student agency as a core value. When AI arrived in schools, most districts reached for the stoplight model. Red means no, green means go, yellow means use caution. It is clear, it is simple, and it is familiar. Ryan understood the appeal. He just didn't think it was enough.
"I don't want kids to go to their jobs and ask if their task falls in red, green, or yellow," he commented. “A stoplight model sets a ceiling on the conversation. Once you know what's allowed, there's nothing left to think about."
As he started thinking about how AI integration should align to Mendon-Upton values and priorities, he designed a framework that moves academic integrity from a disciplinary conversation to an instructional one, and puts the thinking back in the hands of students and teachers.
The Idea Ownership Spectrum
Ryan adapted the Spectrum from a graphic developed by Greg Kulowiec of the Kulowiec Group that addressed ownership and energy level in AI use (MA Department of Elementary and Secondary Education AI Literacy for Educators), then layered in Mendon-Upton’s deeper learning framework and his own thinking about what responsible use requires.
The Idea Ownership Spectrum is an interactive online resource (currently in development) that maps AI use across a continuum of idea ownership. At each level, the framework names both the human role and the AI role, explicitly defining what the learner is responsible for bringing to an idea or task, and what the tool is doing on their behalf.
The levels progress from task completion, where AI is doing more and the human is directing, through drafting and producing, feedback and revision, and knowledge building, up to reflection and synthesis, where the human is doing the deepest thinking and AI is serving as a thought partner or sounding board. The further along the continuum, the more AI literacy, content knowledge, and deeper thinking skills a student or teacher needs in order to use the tool well.

Using the Spectrum
The most concrete application of the tool is what Ryan calls "setting the floor" for a given assignment or project. A teacher opens the tool, toggles off parts of the spectrum that are not appropriate for this particular task, and students can see exactly where their original thinking has to begin. The floor changes based on the learning objective, student content knowledge, the point in the unit, and what the teacher is trying to assess. A research paper might set the floor at drafting. A class discussion might set it to reflection. The goal is to make the pedagogical decision visible and clarify the skills that students must bring to the table.
The most powerful application goes a step further. Instead of the teacher setting the floor alone, the tool can be used during a class discussion, a department meeting, or a coaching session to decide together which level of idea ownership is appropriate for a project or assignment. Students and educators take ownership of the conversation about academic integrity rather than receiving it as a verdict. That shift, from compliance to co-creation, is the heart of what Ryan is building.
Stakeholder Feedback
Ryan is candid that the Spectrum is not a finished product, and the listening sessions he held at the high school and middle school at the end of the year taught him as much as they surfaced for teachers.
Some teachers were engaged and curious. Others approached the initiative with hesitation rooted in past technology integrations where educators felt they had less input, leading to fatigue around rapid, top-down changes. Additionally, initial feedback highlighted anxieties regarding student accountability, reflecting caution about how much autonomy students should have in using AI tools. A growth mindset and a strong commitment to student agency are key requirements for this framework to succeed. Ultimately, working with educators to deepen trust in students is part of the collaboration around developing AI literacy. When teachers and students are equipped with robust content and AI literacy skills, they can safely co-author learning.
Ryan also heard from teachers who couldn't yet imagine the possibilities. "They can't get past what their experience is with the chatbot," he says, "which is: I talk to it and it tells me something." For those teachers, the upper levels of the Spectrum, where AI becomes a thought partner, a reflection tool, or a feedback mechanism, require a level of AI literacy they haven't yet developed.
This points to an important open question that the Spectrum surfaces: the upper levels require AI literacy, content knowledge, and thinking skills that develop over time. A student who hasn't yet built a strong knowledge base in a subject shouldn’t be directing an AI tool to support knowledge building or synthesis in that domain. The Spectrum, used well, becomes a map not just of what AI can do but of what skills students need to demonstrate in order to access those capabilities. That framing applies equally to teachers, many of whom are at the beginning of their own AI literacy journey. Building that literacy, for both students and educators, is a prerequisite for effectively using the Spectrum.
Ryan is still working through structural questions. He wants the tool to flex rather than prescribe, so departments can weigh different variables when making pedagogical decisions about AI use. And he is considering whether the student-facing version of the Spectrum should look different from the educator-facing one.
In the analogy he offers, the framework is the algorithm. The team is still figuring out how to define the product.
Conditions for Success
Mendon-Upton Regional School District serves two towns and operates on a tight budget. Ryan is the primary driver of AI work in the district, with no dedicated instructional technology team below his position. A few important conditions made this work possible.
A human-centered instructional framework was already in place. The deeper learning framework, with its emphasis on student agency, gave the Idea Ownership Spectrum a home. Ryan didn't have to argue for the underlying values. He just had to show how AI fit into them.
Dr. Maureen Cohen, Superintendent of Mendon-Upton Regional School District, was open and willing to engage with the work and contemplate where it was headed, even in a year of significant budget pressure. As an active member of the PEA2K cohort, Dr. Cohen pushed Ryan and the group to put the student and educator experience at the forefront; creating a framework and resources that would easily integrate into classroom practices.
Ryan had the PEA2K cohort. He credits the cross-district conversations for pushing his thinking, particularly the chance to have the framework tested against different contexts and concerns.
Recommendations
Start with responsible use, not acceptable use. Acceptable use tells students what to do. Responsible use asks them to think about why. If the goal is to develop students who can navigate AI well over time, the framework has to build judgment instead of compliance.
Move academic integrity conversations from discipline to pedagogy. The Idea Ownership Spectrum reframes the conversation. Instead of asking whether a student broke a rule, it asks what level of idea ownership is appropriate for this task, and who gets to decide. That is a fundamentally different and more productive conversation.
Use it as a calibration tool. The most powerful application of the Spectrum is a conversation starter at the unit or assignment level. What knowledge and skills do students bring to the task? What are we asking students to learn? What is the AI doing to support this learning? Questions like these are worth asking regularly.
Pair the framework with AI literacy. The upper levels of the Spectrum require students and teachers to know how to direct AI tools intentionally. That literacy needs to be explicitly taught before expecting teachers and students to work at the level of knowledge building or reflection.
Expect the conversation to be harder than the framework. The Spectrum is a starting point. The harder work is the listening sessions, the department meetings, the one-on-one conversations with teachers who have concerns. Plan for that work explicitly.
Invite students into the conversation. The co-creation use case is the most powerful expression of this framework. When students and teachers determine together “the floor” for a given assignment, academic integrity becomes something students own.
As Ryan puts it: "It's fun, and we're not done."
This case study was developed through an interview with Ryan Robidoux, Director of Learning and Innovation at Mendon-Upton Regional School District, in June 2026. It was prepared by Cathy Sanford, Director of Communications at Throughline Learning as part of PEA2K (Partnership for Educators Advancing AI Knowledge) Cohort 1 documentation.
