Skip to main content
Enterprise AI Analysis: Configuring and Monitoring Students' Interactions with Generative AI Tools: Supporting Teacher Autonomy

Enterprise AI Analysis

Configuring and Monitoring Students' Interactions with Generative AI Tools: Supporting Teacher Autonomy

The paper addresses the challenge of maintaining teacher autonomy in education with the widespread use of GenAI tools like ChatGPT. It proposes a middleware system to allow teachers to monitor student interactions, align GenAI outputs with learning objectives, and control system behavior. An initial prototype evaluation with 8 secondary-school teachers showed high perceived usefulness for monitoring student interactions, alerting teachers to issues (e.g., copy-paste), and controlling GenAI outputs. While most teachers perceived higher autonomy, some did not. The study highlights the need for continued refinement, better analytics visualization, and student involvement in system development, while also considering privacy concerns and the potential for institution-specific LLMs.

Executive Impact: Key Insights & Opportunities

Understand the core findings and their direct implications for your enterprise. Each insight presents a tangible opportunity for strategic AI integration.

0 Students using GenAI for assignments
0 Teachers concerned about ChatGPT cheating

Deep Analysis & Enterprise Applications

Select a topic to dive deeper, then explore the specific findings from the research, rebuilt as interactive, enterprise-focused modules.

Teacher Autonomy
Learning Analytics
Human-Centered Design

The paper directly addresses challenges to teacher autonomy posed by GenAI. It defines autonomy as the willingness, capacity, and freedom to control teaching and learning. The proposed system aims to restore and enhance this autonomy by providing monitoring, configuration, and alerting capabilities to teachers.

A core feature of the proposed system is its learning analytics capabilities. It collects and stores student prompts and GenAI answers, allowing teachers to monitor interactions, identify suspicious behaviors (e.g., copy-paste), and analyze recurrently queried topics. This data informs teachers to take action and provide personalized feedback.

The research follows a Systems Development Research Methodology, emphasizing an iterative process that includes user evaluation and feedback. The initial prototype was evaluated by teachers to gather requirements and usability features, aligning with human-centered design principles to ensure the system is adapted to teachers' needs and practices.

89% Students using GenAI for assignments

GenAI tools hinder teacher autonomy by limiting control over student actions and learning processes. Outputs often lack contextualization (e.g., curriculum, student age), impacting course goals. Teachers are unaware of student prompts, GenAI responses, or potential plagiarism. 89% of students used GenAI for home assignments, highlighting a critical gap in teacher oversight and control.

Proposed Middleware System

A middleware system is proposed to mediate between GenAI interfaces and back-ends. It enables teachers to monitor student interactions, align GenAI answers with learning objectives, and configure system behavior (e.g., prompt add-ons, forced hallucinations).

The system follows a Systems Development Research Methodology, with an initial prototype developed and evaluated. This iterative approach ensures the tool evolves to meet real-world teaching needs and feedback.

Design Requirements & Architecture

Key requirements include monitoring student use, alerting teachers to take action (e.g., suspicious copy-paste, recurrent topics), and automatic reactions via teacher configuration (e.g., guidance instead of direct answers, prompt add-ons). The architecture integrates with LMS using LTI, offers teacher and student interfaces, and uses GenAI adapters for multi-tool compatibility.

Construct Conceptual Framework
Develop System Architecture
Analyze & Design System
Build (prototype) System
Observe & Evaluate System
72% Teachers concerned about ChatGPT cheating

Evaluation with 8 secondary-school teachers showed high perceived usefulness for monitoring student interactions, alerting teachers (e.g., copy-paste behaviors), and controlling GenAI outputs. Teachers perceived higher autonomy in configuration scenarios. 72% of surveyed teachers are concerned about the impact of ChatGPT on cheating, underscoring the system's relevance.

Teacher Autonomy Perception

While most teachers perceived higher autonomy, some did not, particularly in monitoring scenarios. Positive feedback highlighted the system's ability to 'guide students in the thinking process' and its 'well-defined and controllable' nature.

Negative feedback pointed to analytics not immediately translating to autonomy or observed classroom work, and the additional time required for setup and monitoring.

Future Directions & Concerns

Future iterations will involve more teachers, use validated questionnaires, and triangulate data. Key concerns include better analytics visualization, involving students in development (due to preference for external tools), data privacy, and developing custom LLMs for curriculum accuracy and data retention.

Aspect Current GenAI Tools Proposed System
Teacher Autonomy
  • Limited control over student interactions
  • Lack of contextualization
  • Unaware of prompts/answers
  • Monitors student use
  • Alerts for suspicious behavior
  • Configurable output (guidance, add-ons)
  • Integrates with LMS
Student Engagement
  • Potential for offloading assignments
  • Copy-paste issues
  • Lack of critical thinking
  • Fosters critical thinking via scaffolding
  • Detects copy-paste
  • Aligns answers with learning objectives
Output Quality
  • Risk of hallucinations
  • Generic, uncontextualized answers
  • No citation
  • Teacher-configured output constraints (e.g., length, forced hallucinations)
  • Contextualized answers (e.g., age-appropriate)

Unlock Your AI Potential: Calculate Your Enterprise ROI

Input your organization's details to see a personalized estimate of potential annual savings and reclaimed human hours with our AI solutions.

Annual Savings $0
Hours Reclaimed Annually 0

Our Proven Implementation Roadmap

A structured approach ensures seamless integration and maximum value realization. We guide you through every phase, from strategy to scale.

Phase 1: Conceptualization & Prototype Development (2-3 Months)

Define research questions, gather requirements, design architecture, and build initial prototype based on Systems Development Research Methodology.

Phase 2: Teacher Evaluation & Feedback Loop (1-2 Months)

Conduct workshops with teachers to evaluate prototype, collect feedback on usefulness, usability, and autonomy perception, and identify refinement needs.

Phase 3: System Refinement & Advanced Features (3-4 Months)

Implement improvements based on feedback, enhance analytics visualization, explore institution-specific LLM integration, and address data privacy considerations.

Phase 4: Student Involvement & Wider Pilot (2-3 Months)

Involve students in co-design, conduct wider pilot studies across different educational levels, and integrate construct-validated questionnaires for broader data collection.

Ready to Empower Your Teachers with AI?

Join leading educational institutions in leveraging intelligent systems to enhance teaching autonomy and student learning outcomes.

Ready to Get Started?

Book Your Free Consultation.

Let's Discuss Your AI Strategy!

Lets Discuss Your Needs


AI Consultation Booking