[Proposal] Course App Category - Mentoring
Overview
We aim to define and streamline default course boundaries for AI integrations by introducing a standard course application category called "Mentoring." This category is designed to encompass a wide range of AI tools that support students within the course experience. The "Mentoring" category builds on the existing Discussions and Live app categories, offering a framework to showcase a growing library of integration options over time.
By establishing standard API expectations, settings, and tools, this proposal seeks to increase visibility and accessibility of existing integrations for more Open edX instances through simplified configuration and directory options in the future.
For the learner experience, any Mentoring app category integration would be able to display interfaces in the learning MFE through various placements, including:
In-context sidebars, akin to existing discussion tools.
Between content blocks within the course.
As standalone pages, similar to Live and Discussions apps.
Additional frontend plugin slots to be determined in the future.
For the educator experience, this proposal envisions that the Mentoring application category will:
Surface existing mentoring tools, such as chatbots, AI sidebar tools, and other student support tools.
Introduce standard and centralized mechanisms to provide per-content XBlock AI mentor prompt guidance and guardrails, ensuring custom pedagogical needs are met and understood by mentoring integrations.
More broadly, this proposal aspires to foster shared ecosystem development speed, enhancing AI course mentoring applications even as tools, models, and underlying technologies evolve rapidly.
Use Cases
AI Chatbot Mentoring: A course integrates a chatbot to provide students with real-time, context-sensitive guidance on assignments and concepts.
Peer Mentorship Tools: Enable students to connect and collaborate with peers in a guided format, facilitated by structured AI prompts.
Personalized Feedback Mechanisms: Allow instructors to deploy AI tools that deliver detailed, personalized feedback to learners.
Career Coaching Integration: Courses with professional development goals could use third-party AI tools for career advice and interview preparation.
Language Practice Conversations: Integrating conversational AI to help learners practice language skills within relevant course contexts.
Goals
Enable seamless integration of mentoring tools via a standardized mechanism.
Expand the scope of interactive and AI-driven learning within courses.
Enhance user experience for both course creators and learners through easy-to-use mentoring integrations.
Foster engagement by providing diverse, personalized, and adaptive mentoring options.
Experience Details
Structure of the Mentoring App Category :
The "Mentoring" section in Pages & Resources would appear alongside existing options like Discussions and Live.
This section would serve as a listing of approved mentoring tools and extensions. How we can allow existing mentoring tools including AI chat bots and more to be added as a mentoring app option should be streamlined to gather existing tools into this area.
Configuration options would allow course creators to set tool-specific parameters such as integration keys, course context, and interaction settings.
Educator Configuration Experience:
Educators access the configurable Mentoring tool from the Pages & Resources card named "Mentoring."
Settings include enabling/disabling where the mentoring tool should render:
Sidebar of content pages (similar to in-context discussions).
Full-page application (similar to Live and Discussions).
Between the content flow of courses.
Educators can add supplemental prompt information or metadata to specific content blocks, guiding mentoring tools to provide structured feedback tailored to those blocks.
Student Experience:
Mentoring blocks can render in three primary ways:
Sidebar of content pages: Mirrors the current in-context discussion model.
Full-page app: Functions similarly to existing Live and Discussions apps.
Embedded in content flow: Appears directly between sections of course content.
These options offer flexibility to ensure that mentoring tools integrate seamlessly into the learning experience.
Technical Considerations:
Standardized APIs for tool integration, similar to existing app categories.
Clear documentation and templates for third-party developers to create compatible tools.
Security measures to ensure data privacy and compliance with educational standards.
UI/UX Enhancements:
Intuitive onboarding and setup flow for instructors.
Modular design to accommodate a variety of tools.
Documented ways to emit analytics from an integration into a platform instances data pipeline to measure tool effectiveness and learner engagement.
Visual concept sketches coming soon
Implementation
Phase 1: Mentoring App Category Basics
This phase aims to develop the foundational framework for the Mentoring area in Studio. We need to create a standardized API structure to enable seamless integration of mentoring tools. This phase should Implement the basic interface for educators to add and configure mentoring tools in the Pages & Resources section.
To start, we could focus on initial support for simple chatbot integrations, with volunteers from the community to be a part of the initial exploration phases to link their existing tools as mentoring apps.
Phase 2: App Category Customization & Safety / Privacy Guardrails
This phase should Introduce customizable configurations for educators, including metadata prompts and guardrails for course-specific content. Whether any of these prompt guardrails or support context can be configured generally is to be determined, but could help ensure safety and privacy based requirements for these tools.
As with the Discussion app category, we would like to consider the feature list table that might be shared across mentoring app categories to help highlight key safety / privacy guardrails in whatever form makes sense for Mentoring - an inclusive and open ended app category.
We should ideally provide templates and developer guidelines to encourage the creation of custom mentoring tools by third-party developers. (This has been split out into Phase 3).
Phase 3: Demo Mentoring App / Starter Tools
This phase aims to develop a proof-of-concept or demo mentoring application hosted in a GitHub repository. This repository would act as a reference or starter code for developers building mentoring applications on Open edX.
Some potential features of this starter repository includes:
A demonstration of the installation or discovery mechanism for this mentoring application, including how a configuration file in the GitHub repository might provide an Open edX instance with a simple installation process via the GitHub URL or another straightforward mechanism.
Potential demo configurations for integrating ChatGPT or similar tools.
Sample API implementations showcasing standard mentoring workflows.
Links to documentation to guide developers in adapting the code to their own mentoring tools.
Demonstrate how this metadata would integrate seamlessly into the educator configuration experience, aligning with the standard metadata structure used for Discussions and Live applications in Open edX.
Provide educators with the ability to test this demo application within their courses, offering insights into how mentoring tools can enhance the learner experience.
Use feedback from developers and educators to refine the repository and improve mentoring tool development processes.
Draft: Mentoring App Category Metadata
Metadata setup to include:
Application title, image, short description, and long description.
Links to key documentation for the application.
Information about the application maintainer, including contact details for technical support.
Phase 4: Expanded Mentoring App Integration List & Iteration
After initial delivery of Phases 1-3, w would like to refine the mentoring integration experience based on pilot feedback, improving the interface and addressing usability concerns.
We could consider exploring how analytics could be shared back into the platform for educators to monitor learner engagement and tool effectiveness.
If this hasn’t been completed previously we need to ensure the Mentoring App category is prepared for broader platform rollout with comprehensive documentation and support.
Feedback and Iteration
[Details to be filled based on the feedback mechanisms planned.]
Alternatives Considered
Direct Integration Without a Standardized Area:
Pros: Simplifies immediate implementation.
Cons: Lacks scalability and standardization, leading to inconsistent user experiences.
Risks and Mitigations
Risk: Complexity in integrating diverse mentoring tools.
Mitigation: Provide comprehensive documentation and developer support.
Risk: User adoption may be slow due to unfamiliarity with new tools.
Mitigation: Design intuitive interfaces and provide training resources for instructors.
Estimated Effort
[Details to be filled based on development resource planning.]
Conclusion
The addition of a "Mentoring" area in the Pages & Resources section of Studio represents a significant opportunity to expand the Open edX platform's capabilities. By enabling standardized, scalable integrations of AI-driven mentoring tools, this feature aligns with the mission to enhance learning experiences and meet the evolving needs of educators and learners.
Appendix
Other Known Open edX, AI, and Chat-Related Integrations:
Tutor-ChatGPT Plugin:
A plugin designed for Open edX's Tutor environment to enable ChatGPT-based interactions for learners within courses. This integration provides a simple way to offer conversational AI features using OpenAI's API.
LangChain for Open edX:
A flexible library that supports embedding large language models (LLMs) into Open edX environments for tasks such as conversational tutoring, summarization, and guided learning paths.
Rasa Open Source:
An open-source conversational AI framework that can be integrated into Open edX to build custom chatbots tailored to course-specific needs.
Dialogflow for Learning Platforms:
A natural language understanding platform by Google that can be used to create conversational agents for Open edX courses, offering automated help and contextual support.
Anthropic Claude Integration:
Examples of using Anthropic's Claude model in educational environments for AI-driven assistance, which could be adapted to Open edX use cases for mentoring and guidance.
DeepPavlov AI Assistants:
An open-source framework for building dialogue systems, adaptable for educational platforms like Open edX to provide course-level support.
This section highlights notable tools and frameworks that have been explored or used in conjunction with Open edX, AI, and chat-based features, offering inspiration and guidance for future integrations.