A faculty-focused tool for evaluating how easily generative AI can satisfy an assignment and then guides you through redesigning assessment so students still have to demonstrate the skills that matter most to your assessment.
Open the AI-Proofing Assistant →
No Amazon Quick account or coding experience is required to use the public app.
The AI-Proofing Assistant helps faculty evaluate an assignment or rubric for vulnerability to AI-generated submissions.
Provide your assignment goals, learning objectives, instructions, and grading criteria, then define the maximum grade you believe an AI-only submission should be able to earn. The app generates:
- an AI Robustness Score
- a SWOT analysis of the current assessment
- three redesign approaches: Minimum Robustness, Heightened Security, and Best Fit
This app identifies where AI is a likely substitute for the skills an assignment is intended to assess and where requirements can be redesigned to emphasize students' skills demonstration.
Designed for: Faculty, instructional designers, faculty developers, teaching and learning centers, and educators exploring responsible AI use in assessment.
- Open the app.
- Enter or upload your assignment instructions, objectives, grading criteria, or rubric.
- Set the maximum grade you want an AI-only submission to be capable of earning.
- Review the AI Robustness Score and SWOT analysis.
- Compare the Minimum Robustness, Heightened Security, and Best Fit redesign options.
- Export useful recommendations and adapt them to your assignment or LMS.
Supported uploads: .txt, .md, .csv, .json, and .rtf, up to 5 MB.
- AI Robustness Analysis: Estimates how successfully generative AI could satisfy the existing assignment requirements.
- SWOT-Based Review: Identifies where the assessment already requires authentic engagement and where AI may substitute for student performance.
- Faculty-Defined Grade Ceiling: Uses your target maximum AI-only grade to guide recommendations.
- Three Redesign Levels: Offers options ranging from small rubric adjustments to more substantial changes in instructions, evidence, or submission requirements.
- Exportable Recommendations: Produces revisions that can be further edited and incorporated into course materials.
Getting Started FOR FREE in Amazon Quick A one-minute introduction to accessing the Amazon Quick environment.
AI-Proofing Your Assignments: AWS QuickDive See the full workflow from assignment input through robustness scoring, SWOT analysis, and practical redesign recommendations.
Build an AI-Enhanced App Without Coding: Creating with Amazon Quick Follow the creation of an AI-powered application using natural-language development, iteration, troubleshooting, versioning, customization, and publishing in Amazon Quick.
View the complete video playlist →
Want to see what the process looks like before using the app?
A documented build example showing how the application and its analysis workflow were developed.
A completed analysis demonstrating how the tool identifies AI vulnerabilities and recommends changes to an existing applied assignment.
A second completed example showing how recommendations change when the assignment goals, evidence requirements, and structure differ.
📸 Want a visual preview? Browse the app screenshots and project images to see the Assignment Defender interface and visual layout for the example outputs.
These freely available resources provide a pathway from understanding generative AI to developing AI-enabled applications.
AWS Machine Learning University - Generative AI Learning Path Generative AI fundamentals, responsible AI, prompt engineering, foundation models, RAG, agents, multimodal systems, and application development with Amazon Bedrock.
AWS MLU - Agentic AI Learning Materials Continue into agentic systems that reason through multistep tasks, use tools, and participate directly in application workflows.
AWS MLU Generative AI - Module 3: Building Applications with Foundation Models Focus specifically on conversational applications, RAG, agents, and multimodal application development.
Amazon Quick Documentation Official documentation for creating, iterating, testing, publishing, and sharing applications in Amazon Quick.
Do I need an AWS account or coding experience?
No. The public application can be opened from the shared link without an Amazon Quick account, and using the app requires no programming experience. An account becomes relevant if you want to create or duplicate applications yourself.
What is a SWOT analysis?
A structured way of assessing and communicating levels of risk and security. Strengths and Weaknesses draw from historical and current data, while Opportunities and Threats address future potential and forecasting.
What do the three redesign levels mean?
Minimum Robustness emphasizes smaller changes, particularly to rubric criteria and weighting.
Heightened Security adds stronger changes to assignment instructions, evidence, or submission requirements.
Best Fit seeks the strongest practical combination of instructions, requirements, and grading criteria while preserving the assignment's central goals and learning objectives.
Does this make an assignment completely “AI-proof”?
No. The AI Robustness Score is an estimate intended to support instructional decision-making. AI capabilities continue to change, and no assessment design can guarantee that AI will not be used.
The goal is to identify where AI can substitute for the skills being assessed and strengthen requirements that call for genuine student performance.
Is the goal to prevent students from using AI?
Not necessarily. Appropriate AI use may be allowed or encouraged while an assessment still requires students to demonstrate judgment, evidence use, interaction, reflection, decision-making, or other intended learning outcomes.
The AI-Proofing Assistant is a formative instructional-design and decision-support tool. Its scores, SWOT analyses, and redesign recommendations intended to inform faculty judgment.
Do not submit student names, IDs, identifiable grades, education records, or other confidential, restricted, personally identifiable, or FERPA-protected information. The application is intended to analyze assignments, learning objectives, instructions, rubrics, and related instructional materials, not individual student records.
Output quality depends on the specificity of the information provided. Recommendations should be reviewed against course learning outcomes, accessibility needs, disciplinary expectations, workload, and institutional policies before implementation.
Last tested: September 2026
Educational and documentation content is licensed under CC BY 4.0.
Developed by Nathan Kelly, Professor of Political Science, Oklahoma City Community College, through work focused on generative and agentic AI, assessment design, and faculty workforce upskilling.
This project draws on freely available learning and technical resources from AWS Machine Learning University (MLU) and Amazon Web Services. AWS and Amazon product names and trademarks remain the property of their respective owners.
This repository represents independently developed educational material and is not an official AWS product, endorsement, or institutional policy statement.
Contact: Nathan Kelly Professor of Political Science, Oklahoma City Community College AWS EEP MLU 2026 Faculty · AWS EEP Faculty Fellow 2026–2027 nathan.d.kelly@occc.edu
Built with ❤️ using Amazon Quick