This epic covers the discovery and initial delivery of an AI subsystem in Moodle. 

An AI Subsystem in Moodle LMS will create a way to provide a consistent and user-friendly way for users to interact with AI in Moodle's user interface, as they do their teaching and learning activities. As well as providing a straightforward way to integrate with various AI providers on the backend.

Moodle AI forums and survey feedback will be included as part of the design and delivery process; in addition to internal discovery activities. There will also likely be activities conducted via the Moodle user experience lab. Reviewing existing API designs of popular AI services and related work will also be investigated.

These and related activities will be collated and communicated via this epic. 

Discovery and implementation of a solution will be guided by and adhere to our AI Principles especially Transparency, Configurability and data protection.

The linked issues to this epic will cover the individual design and implementation activities. There will be a focus on prototyping and iterative design, allowing short feedback cycles and adjustments.

This work will not cover the creation of an AI model (LLM, Generative model, etc.) only the integration of models. Although, an open source model may be included in future iterations of development. 

Initial work in MDL-78113 will be used as a reference for a technical approach.

Linked child issues contain further information about components and approach.{}

Approach update:

2024-05-16: Below is a short update video on our current approach and outline of initial deliverables.

AI_update_1.mp4

Design Overview: