Skip to Main Content

notice Some items are only available on campus or will require authentication via EUID and Password at the point of use.

Assessment @ UNT Libraries

This guide is meant as an introduction to assessment at the UNT Libraries with relevant resources for anyone interested in starting their own assessment project.

Artificial Intelligence (AI) in Assessment

These are a summary of notes from a guided discussion about using AI in assessment from the UNT Library Assessment Committee on October 23, 2025.

Disclaimer: This is not a definitive guide. This is an evolving topic, and we wanted to add notes from our discussion to help share what we have explored and learned as well as what we see as hurdles.

What have we used LLM’s for?

  • It is good at things that are formulaic and don’t require very specific, niche knowledge.
  • Create learning objectives and rubrics
  • Make goals into SMART goals
  • Used for qualitative coding, both to help save time and to remove bias
  • Helped with ideation and starting a project that feels insurmountable
  • Need a human in the loop to review results and give it a human touch
  • Is good for first drafts and project ideas

When is GenAI OK to use?

  • Make sure to credit the specific tool that you use.
  • Always review the things that you generate.
  • If you would feel comfortable having it clearly and openly credited, then it should be OK to use.

How do we deal with the lack of reproducibility?

  • If you run the same analysis with the same data and same prompt twice, you will get two different results.
  • This isn’t just a concern for us – it’s something that the greater area of scholarly communication is going to have to grapple with.
  • Keep a log of the steps in your analysis and all of your prompts.
  • Report your prompts and what LLM model you used as part of your methodology.