What is this course about?
AI detectors produce false positives that land disproportionately on multilingual students, and honor-code-only responses do not scale. This course starts from the evidence on detection, then rebuilds assessment around evidence AI cannot manufacture. Learners leave having converted one real high-enrollment assignment and drafted a syllabus clause that survives a grade appeal.
Who is this course for?
- Faculty across disciplines
- Department chairs and curriculum committees
- Centers for teaching and learning staff
- Academic integrity officers
What do I need before starting?
- Bring one real assignment you currently give
- No technical background required
What will I be able to do afterwards?
- State what the evidence does and does not support about AI detection
- Redesign an assignment so that AI use does not undermine its evidence of learning
- Build process-based assessment that scales to a large section
- Run oral defense checkpoints without an unsustainable time cost
- Write a syllabus AI clause that holds up in a grade appeal
What does each module cover?
What does the evidence say about AI detectors?
40 minThe false positive problem, who it lands on, and why detector output cannot carry a grade appeal.
Objectives
- Summarize the evidence on detector accuracy and bias
- Explain why base rates make even a good detector unusable at scale
- Determine what role, if any, detection output should play
Topics
Activity. Work through the base-rate arithmetic for your own section size and see what a positive result means.
Why does an honor-code-only response fail?
35 minThe limits of policy without design change, and the equity consequences of selective enforcement.
Objectives
- Identify where policy-only responses break down
- Describe the enforcement asymmetry that policy alone creates
- Make the case for redesign to a skeptical department
Topics
Activity. Draft the three-minute case for redesign you would make in a department meeting.
How do you assess process rather than product?
50 minDrafts, version history, and revision narratives as primary evidence of learning.
Objectives
- Design an assignment where the process is the graded artifact
- Use version history as evidence without surveilling students
- Grade revision narratives efficiently
Topics
Activity. Convert one assignment to a process-graded structure with a revision narrative requirement.
How do you run oral defense at scale?
50 minViva-style checkpoints in large sections — the format that most reliably distinguishes understanding from output.
Objectives
- Design a three-minute defense protocol
- Schedule defenses in a large section without consuming a term
- Score consistently across multiple graders
Topics
Activity. Design and time a three-minute defense protocol, then run it with a colleague.
What assignments require your local context?
45 minBuilding assessment around data, materials, and settings that no general model has access to.
Objectives
- Identify local context that is genuinely inaccessible to a model
- Design assignments requiring primary local engagement
- Balance local specificity against transferable learning
Topics
Activity. Redesign one assignment around a local data source or setting.
How do you design assignments that require AI use?
45 minPermitted-use design where AI is a required, cited, critiqued tool rather than a prohibited one.
Objectives
- Design an assignment where AI use is required and evaluated
- Require citation and critique of AI output
- Assess the student's judgment about the tool
Topics
Activity. Build a permitted-use assignment where the student must find and correct the model's errors.
What syllabus clause survives a grade appeal?
40 minWriting an enforceable AI policy — specific, proportionate, and consistent with institutional policy.
Objectives
- Draft assignment-specific rather than blanket AI policy
- Align the clause with institutional academic integrity policy
- Anticipate the appeal arguments the clause must answer
Topics
Activity. Write your syllabus AI clause and have a colleague attack it as an appealing student would.
Converting a high-enrollment assignment end to end
60 minThe workshop module: a complete conversion with rubric, timeline, and grading load estimate.
Objectives
- Produce a fully converted assignment with a rubric
- Estimate the realistic grading load
- Plan the communication to students
Topics
Activity. Complete the conversion and present it to the cohort for critique.
What is the capstone project?
Department assessment redesign package
Convert one high-enrollment assignment completely, produce the rubric and grading load estimate, draft the syllabus clause, and package it so colleagues in the department can adopt it directly.
Deliverable: A department-adoptable assignment package with rubric, syllabus language, and student-facing explanation.
How are learners assessed?
- Peer critique of the converted assignment against a redesign rubric
- Syllabus clause stress-tested by a colleague playing an appealing student
- Grading load estimate checked for realism against actual section size
What ships with the course?
Facilitator guide
Session-by-session running order, discussion prompts, and the questions that reliably derail a room.
Learner workbook
Exercises, checklists, and the templates each module's activity produces.
Hands-on lab environment
A sandboxed ibl.ai deployment so exercises run against real agents, not screenshots.
Assessment bank
Scenario questions and rubric criteria mapped to each stated learning outcome.
Source bibliography
Every primary regulation and standard cited on this page, linked and dated.
Which AI agents does this course use?
The hands-on modules run against agents already deployable on the ibl.ai platform for higher education.
Where does the course material come from?
Every module is grounded in primary sources — the regulation, standard, or research itself, not a summary of it. Each was resolved at authoring time.
- Office of Educational Technology
U.S. Department of Education
Federal framing for AI in teaching, learning, and assessment.
- Copyright and Artificial Intelligence
U.S. Copyright Office
Authorship questions relevant to citation conventions in Module 6.
- AI Index Report
Stanford HAI
Evidence on model capability, used to calibrate what assignments AI can complete.
- AI Risk Management Framework
NIST
Framing for the reliability and bias discussion around detection tools.
Delivery notes
Binding guidance for anyone preparing and delivering this course.
- Module 1's base-rate exercise must use the learner's own section size. Doing the arithmetic themselves is what changes minds — a cited statistic does not.
- Every module needs a worked example from at least three disciplines including a quantitative one. Faculty dismiss this material instantly if every example is an essay.
- Be explicit that the grading load of process-based assessment is real and often higher. A course that pretends otherwise loses the room in Module 3 and never recovers.
- Do not recommend any specific detection product, including favorably. The Module 1 position is that detector output cannot carry an appeal, and a product recommendation contradicts it.
- The oral defense module needs a real timed demonstration. Facilitators consistently underestimate how much three minutes reveals; show it rather than assert it.
Why run AI training on a platform you own?
You own the course, not a licence to it
Course content, learner data, and the platform run inside your perimeter — you own all the code and the data.
Model-agnostic delivery
Run the course's AI components on any LLM — Claude, GPT, Llama, Gemini, Command — and switch anytime.
No per-seat training licences
Usage-based or self-hosted, so cost tracks actual use rather than headcount.
Deploy anywhere
Cloud, private VPC, on-premise, or fully air-gapped — including for cohorts that cannot use public AI tools.
Frequently asked questions
What does the Assessment Redesign for the AI Era course cover?
AI detectors produce false positives that land disproportionately on multilingual students, and honor-code-only responses do not scale. This course starts from the evidence on detection, then rebuilds assessment around evidence AI cannot manufacture. Learners leave having converted one real high-enrollment assignment and drafted a syllabus clause that survives a grade appeal. It runs 5.5 hours across 8 modules across 8 modules, at foundational level, and closes with a capstone: Department assessment redesign package.
Who should take Assessment Redesign for the AI Era?
It is written for Faculty across disciplines, Department chairs and curriculum committees, Centers for teaching and learning staff, Academic integrity officers. Prerequisites: Bring one real assignment you currently give; No technical background required.
Can we run this course on our own infrastructure?
Yes. ibl.ai is model-agnostic and deploy-anywhere — cloud, private VPC, on-premise, or fully air-gapped — and you own all the code and the data. Cohort data, submissions, and any material learners upload stay inside your perimeter, which matters for higher education teams that cannot send work to a public AI tool.
How do we get access to Assessment Redesign for the AI Era?
Request access and we will set it up for your cohort — hosted by ibl.ai, or running against your own deployment. Tell us the group size and timing you need, and whether it should run inside your own perimeter.
How much does AI training for higher education cost on ibl.ai?
There is no per-seat pricing — you pay for usage or self-host and pay only for the infrastructure, so a 5,000-person rollout does not cost 5,000 licences. 1.6M+ users across 400+ organizations run the platform this way, including NVIDIA, MIT, and Syracuse University.