"Education is not only about acquiring information. It is about forming judgment. Artificial intelligence can generate responses. Education must develop thinkers."
-Dr. Lydia Elliot, Carle Illinois College of Medicine at the University of Illinois Urbana-Champaign, 2006
This fall, faculty will likely juggle two kinds of readiness: the exciting work of preparing to teach students and considering generative AI's effect on how we assess what students turn in as evidence of learning. In this brief, we identify early steps faculty can use to plan for successful student engagement and then we share ideas for ways we might reimagine student learning assessment that is informed by the presence of generative AI.
What catches instructors off guard in the first weeks?
Faculty who are new to college teaching — including experienced K–12 educators making the jump — consistently report the same early surprise: the emotional and communication load of the job outweighs the challenge of content delivery, and formal onboarding is often thin (Morales, 2026). Many also arrive expecting students to be highly self-directed, only to find that explicit structure — reminders about deadlines, regular checks for understanding, repeated scaffolding — benefits students at every level, not just first-years (Morales, 2026).
"Find your person," one instructor advised - counsel that matters during a stretch when many new instructors feel professionally isolated (Morales, 2026).
1) Before Day One
- Share information about you and your course through your syllabus. Develop a clear, student-focused guide that consolidates essential course information and tools for navigating the semester. Consider meeting with UCITE for a syllabus session if you'd like to explore ways to craft a document that students can meaningfully use to anchor their learning.
- Confirm your course shell. Publish your Canvas site, and state your grading scheme, attendance policy, and late-work policy in plain language - not just implied.
- Check accessibility. Captioned media, tagged PDFs, and readable contrast on visual materials (i.e. slides) - take a few minutes now and prevent accommodation scrambles later (https://case.edu/accessibility).
- Talk with colleagues in your department. Ask about shared exams, common rubrics, or coordination expectations so your section aligns with others.
2) Your First Weeks
- Open (almost) every class with orientation that sparks academic inquiry. Start class by proposing a question, sharing a quote or an image that provokes student curiosity, or by relating your research to your course context. Follow with a quick rundown of what's due, what's coming, and what today covers goes a long way toward the structure students still need (Morales, 2026).
- Build in regular checks for understanding. Short discussions, low-stakes questions, or a quick poll tell you far more than silence does.
- Expect a steady stream of individual communication. Absences, extensions, and accommodation requests are a normal and significant part of the workload, not a sign that something is going wrong in your course (Morales, 2026).
3) You Are Not Alone
- Find a colleague or mentor early. Your "person" could be a department colleague and/or a colleague outside your discipline that can provide a different CWRU perspective. Your "people" should be folks you can ask about the small logistical questions that no handbook covers.
- UCITE (verb tense)! Engage with CWRU's center for teaching and learning through new-faculty orientation, teaching consultations, and conversational programming. We are here to support you throughout the fall - not just in week one.
How is AI reshaping teaching, learning, and assessment? What should faculty consider as we start a new semester?
A growing number of institutions — including Yale, Vanderbilt, Johns Hopkins, and Indiana University — have restricted or discouraged faculty reliance on AI-detection software as evidence of cheating, and several have disabled Turnitin's AI-detection feature outright (Palmer, 2026). The tools have shown real problems: inconsistent results, false positives on human-written text, and documented bias against multilingual and non-native English writers, with some institutions facing legal challenges from students flagged incorrectly (Palmer, 2026).
At the same time, a deeper issue is emerging alongside the integrity question. Generative AI can produce fluent, correct-looking answers without the interpretation, judgment, or reasoning an assignment was meant to build — which means correct output no longer reliably indicates understanding (Elliott, 2026).
“We are beginning to mistake performance for understanding,” as Elliott (2026) put it, describing the gap between what AI can produce and what students actually comprehend.
None of this is new so much as newly visible. Assessment practices that reward finished output over demonstrated thinking were already fragile; AI has simply made that fragility impossible to ignore (Clark, 2026).
“It is a moment to redesign,” Clark (2026) writes — not to chase every new AI tool with a corresponding detector.
1) Focus on appropriate academic challenge and student supports to meet those challenges (i.e. rigor)
- UCITE hosted Dr. Kevin Gannon as keynote speaker at Teaching Forward, our annual celebration of teaching excellence. He challenged faculty to ask one question when considering the use of any learning tool, “Does this advance student learning?”
- AI Detectors are unreliable enough that several universities now restrict their use as evidence (Palmer, 2026).
2) Evaluate Thinking, Not Just Output
- Add a short oral component to written work: a brief conversation where a student explains or defends their own submission in their own words. If you have a large class, consider a timed video submission or brief presentation (i.e. research lightning talk).
- Use case- or scenario-based prompts that ask students to apply a concept to an unfamiliar situation, not just describe it (Elliott, 2026).
- Ask students to “show their work”: by documenting the learning process and not solely learning products. CWRU Nursing faculty colleagues have redesigned their curriculum to focus on process assessments through competency-based learning.
3) Anchor Assignments to the Specific
- Tie assignments to something that happened in your specific class: a discussion, a dataset, a current event you covered — an AI agent can write a competent generic essay, but not one that references last Thursday's debate (Clark, 2026).
- Incorporate scaffolded (i.e. stepwise) project- and scenario-based work that requires judgment and context. You might start with formative, short-answer assignments that allow students to fail/succeed/learn in low/no-stakes contexts.
4) Make Thinking Visible Over Time
- Ask for the scaffolding, not just the final product: planning notes, outlines, drafts, and annotated sources submitted alongside finished work (Clark, 2026).
- Incorporate reflective journals that build a record of a student's thinking and developing confidence across the semester.
5) Be Present
- Stay actively engaged: substantive feedback, active discussion moderation, and following up individually make patterns in student work visible and remain one of the strongest deterrents to any form of academic dishonesty, AI-assisted or otherwise (Clark, 2026).
Summary
Two commitments run through both halves of this brief:
- Structure is not coddling. Clear expectations, regular check-ins, and visible instructor presence serve students at every level and, as a side effect, make academic dishonesty — AI-assisted or otherwise — harder to hide.
- Design beats detection. Chasing each new AI tool with a corresponding detector is a race we are not positioned to win; anchoring assessment to specific, process-visible, judgment-based work is a more durable fix.
UCITE invites you to bring either set of questions — about your first weeks, or about redesigning a fall assignment — into a conversation with us. Schedule a teaching consultation, bring a draft assignment to a workshop, or just reach out. Visit our website (https://case.edu/ucite/teaching-resources) for more information or contact us: ucite@case.edu.
Download this article in PDF!
Selected References
Clark, L. A. (2026, July 13). AI didn't break assessment: It exposed what was already broken. Faculty Focus.
https://www.facultyfocus.com/articles/educational-assessment/ai-didnt-break-assessment-it-exposed-what-was-alreadybroken/
Elliott, L. (2026, July 31). Why AI is changing what it means to be intelligent. Faculty Focus.
https://www.facultyfocus.com/articles/educational-assessment/why-ai-is-changing-what-it-means-to-be-intelligent/
Morales, C. (2026, August 3). What nobody warns you about teaching college for the first time. Faculty Focus.
https://www.facultyfocus.com/articles/faculty-development/what-nobody-warns-you-about-teaching-college-for-the-firsttime/
Palmer, K. (2026, August 5). AI detectors are out, new approaches are in. Inside Higher Ed.
https://www.insidehighered.com/news/tech-innovation/artificial-intelligence/2026/08/05/ai-detectors-are-out-newapproaches-are