Earlier this fall, the Emerging Technology team within ITS released several improvements and new features for our custom generative AI tool, Maizey. As the academic year unfolds, these features will help faculty, staff, and students to streamline workflows, create and interact with custom AI tutors, and more.
U-M Maizey is now integrated with Canvas, enabling faculty and instructions to easily create custom AI tutors for their courses right within the Canvas interface. This integration streamlines the creation process and in less than ten minutes, instructors can configure a custom AI tutor trained on their course materials, offering students 24/7 access to a chatbot trained on the course content. Hear from faculty who have used Maizey to train an AI tutor in this video.
Platform-Wide Enhancements to Maizey
In addition to the integration with Canvas, Maizey AI has significant upgrades that enhance its capabilities across the entire platform: multiple-source data integration and automated reindexing.
Prior to these upgrades, Maizey was only able to reference one type of data source (i.e. Google Drive OR Dropbox OR Canvas OR external websites). Now, Maizey can reference multiple data sources at once so information stored in any of the approved sources can be used to train the AI, ensuring comprehensive and informed answers to user prompts.
In addition to the ability to reference multiple data sources, users can now schedule Maizey to reindex the data sources it is trained on. This allows Maizey access to the most up to date information from the materials it references. Set Maizey to reindex as frequently as once a week or build a schedule that meets your specific needs.
For more details on using Maizey, visit the support page or get started right away.