AI Enhanced Workflow for Academic Integrity Checking

Introduction

This workflow outlines a comprehensive process for leveraging AI technologies to enhance academic integrity checking for student work. It integrates automated systems and human oversight to efficiently detect and address potential integrity issues while providing timely feedback to students.

Process Workflow for AI-Enhanced Academic Integrity Checking for Student Work

Initial Submission and Processing

1. Students submit their work through a secure online portal integrated with the institution’s learning management system (LMS).
2. An AI-powered document processing system, such as Artificio or Solix ECS, automatically:
   • Classifies the document type (e.g., essay, lab report, code).
   • Extracts key metadata (student ID, course code, assignment details).
   • Performs optical character recognition (OCR) if necessary.
   • Converts files to a standard format for analysis.
3. The processed documents are stored in a secure cloud repository with version control and access logging.

Automated Plagiarism Detection

4. The standardized documents are analyzed using multiple AI-powered plagiarism detection tools in parallel:
   • Turnitin’s AI content checker to identify potential use of AI writing tools such as ChatGPT.
   • iThenticate to compare against published academic sources.
   • A custom-trained machine learning model to detect similarities with the institution’s internal repository of past student work.
5. Results from these tools are aggregated and scored on a unified scale.

AI-Assisted Human Review

6. Documents exceeding a configurable similarity threshold are flagged for human review.
7. An AI assistant, such as Perplexity or Claude, analyzes the flagged sections and generates a summary report for the reviewer, highlighting:
   • Specific passages of concern.
   • Potential original sources.
   • Explanations of why the content may be problematic.
8. Human reviewers (e.g., teaching assistants or instructors) utilize this AI-generated report to efficiently assess potential academic integrity issues.

Automated Feedback Generation

9. For submissions not requiring human review, an AI writing assistant, such as Grammarly or ProWritingAid, generates constructive feedback on:
   • Grammar and style.
   • Citation formatting.
   • Argument structure and clarity.
10. This automated feedback is made available to students immediately through the LMS.

Decision and Documentation

11. For flagged submissions, reviewers document their findings and decisions in a standardized form within the academic integrity management system.
12. An AI-powered decision support tool provides recommendations on appropriate outcomes based on institutional policies and precedent cases.
13. Final decisions and supporting evidence are automatically compiled into a formal report using dynamic document generation.

Communication and Appeals

14. Automated, personalized communications are sent to relevant parties (student, instructor, academic affairs office) based on the outcome.
15. In cases of suspected violations, students can initiate an appeal through an AI-guided interface that assists them in understanding the process and submitting relevant information.
16. Appeal documentation is automatically routed to the appropriate review panel or administrator.

Data Analysis and Reporting

17. All case data is aggregated into a central database for analysis.
18. AI-powered analytics tools generate insights on:
    • Trends in academic integrity issues across courses and departments.
    • Effectiveness of prevention and detection measures.
    • Consistency in outcomes and decision-making.
19. These insights inform policy updates and targeted interventions to promote academic integrity.

Continuous Improvement

20. Machine learning models used throughout the process are regularly retrained on new data to enhance accuracy.
21. Natural language processing algorithms analyze student and faculty feedback to identify areas for workflow optimization.
22. The entire system undergoes periodic audits to ensure compliance with privacy regulations and ethical AI principles.

This AI-enhanced workflow significantly improves efficiency, consistency, and effectiveness in managing academic integrity compared to traditional manual processes. Key benefits include:

• Faster initial screening of submissions.
• More comprehensive and objective plagiarism detection.
• Reduced workload for human reviewers.
• Timely, constructive feedback for students.
• Standardized documentation and decision-making.
• Data-driven insights for policy improvement.

By leveraging multiple AI technologies at each stage, the process balances automation with human judgment to uphold academic standards while supporting student learning and growth.