Building unit-specific AI assistants

Image generated using Flux AI (Dev)

Purpose

Often it's important for AI to be aware of the topics covered in your unit and support student learning within that context. This may be where the AI acts as a Socratic tutor on unit content, or where the AI supports students where you know they need more guidance.

---

Building a custom AI agent

To achieve this, we usually need to build a custom AI agent rather than relying on special prompting in a general-purpose AI tool like Microsoft Copilot Chat or ChatGPT. This is because the custom AI agent can be told what topics to focus on, how it should interact with students, and perhaps be given specific documents that are only available to students through the unit.

This can be done through platforms such as Cogniti. Below are some examples of such unit-specific AI assistants.

 

Example: Dr MattTabolism for biochemistry

This AI agent was built to support second year biochemistry students unpack lecture content in a supportive and Socratic fashion. You can read more about Dr MattTabolism on Teaching@Sydney.

 

Example: Essay topic interlocutor

This AI agent was built to provide students feedback on essay topics and help them develop their English literature skills. The unit coordinator embedded interactions with the AI agent as part of an assessment task to encourage more meaningful engagement with AI. Designing AI agents is as simple as providing plain-language instructions to the AI. Read more about this agent on Teaching@Sydney.

 

Example: Dr HistoPathix for dentistry

This AI agent was built to support dentistry students learn oral histopathology through case scenarios in a Socratic fashion. It provided students with personalised and adaptive learning experiences and challenged students to articulate their thought processes. Watch the academic talk about Dr HistoPathix in the video below.