We help schools replace reactive AI decisions with a coherent approach, grounded in how students actually learn and develop.
We help schools replace reactive AI decisions with a coherent approach. SCAFFOLD gives educators a practical way to decide where AI strengthens learning and where students do the work themselves.
Students are already using AI. Parents and boards want schools to prepare students for a changing world. New AI features keep appearing in the tools schools already use. Teachers are making decisions in real time.
Under that pressure, schools often make decisions out of order. We take a different approach.
Faculty and leaders first build a common understanding of what has actually changed about teaching and learning.
From there, a shared way to think about AI grounded in how students actually learn and develop.
Teachers apply that lens to real assignments, assessments, and the decisions they make every day.
Guidance grows out of the practices you want to support, rather than trying to anticipate every new tool.
Students develop knowledge, skill, and judgment by doing the cognitive work themselves. AI can support that work or do it for them. The same tool does either, depending on the learning goal, the student, and how the interaction is designed.
The goal is not to push every use of AI toward one end of the spectrum. It is to make deliberate choices about which cognitive work students need to do themselves and where support can help them develop.
A useful scaffold changes as competence grows. The same tool shouldn't look the same in 7th grade and 11th grade, and these eight principles help teachers make that call for their own students.
What should students be able to understand or do by the end? Everything else follows from this.
Give AI a purpose that serves the goal: tutor, coach, questioner, critic, or practice partner.
Design interactions that ask AI to question, quiz, or give feedback rather than produce the finished work.
Build students' ability to evaluate AI output, check sources, and decide what is worth using.
Make student thinking visible through drafts, checkpoints, and explanations along the way.
Students should be able to explain, defend, and extend the work they submit.
Create transparency around how AI contributed: the tools used, key prompts, and what changed.
Include a moment where students independently show what they understand, through defense, reflection, or application.
(S · C · A) Start with what students need to learn, then design any use of AI around the thinking and practice required to get there.
(F · F) Help students evaluate AI critically, monitor their own understanding, and recognize when AI is supporting their thinking and when it is beginning to replace it.
(O · L · D) Build accountability and evidence throughout the process so teachers can see what students understand, how they got there, and what they can do independently.
Teachers design assignments and assessments that reveal what students actually know, understand, and can do, even when AI is readily available outside the classroom.
Departments work from shared principles while preserving the professional judgment teachers need across different disciplines, ages, and learning goals. Students feel the consistency.
Instead of blanket bans or broad permission, faculty decide where AI strengthens the learning goal, where it gets in the way, and where students do the work themselves.
Schoolwide guidance grows out of the teaching practices your school actually wants to support, rather than trying to anticipate every new tool or use case.
Expectations for AI are clearer because the school can explain the learning rationale behind them to students and parents, not simply what is permitted or prohibited.
Tools will change. A framework grounded in learning and development gives educators a way to keep making thoughtful decisions as they do.
Intrepid Scholar helps K-12 schools rethink teaching, learning, and assessment for a world with AI. Our work combines developmental psychology, cognitive science, AI fluency, and practical implementation.