AI: UT AI Hub Frequently Asked Questions (FAQs)

I have a professor paying for a monthly Claude Pro subscription. Do you know if the Claude Pro will cover different tiers as they offer for upgrades?

Claude Sonnent 4.6 and Claude Opus 4.6 will both be available as part of the UT AI Hub

Could you please clarify the following technical specifications regarding the university's Claude deployment?

  1. Backend Architecture: Is the access routed through a pay-as-you-go API (which removes standard session limits) or a seat-based enterprise/team license?
  2. Usage Limits: If it is seat-based, what are the specific message, token, or hourly limits enforced per user?
  3. Timeout Protocol: Does the university's interface include a visible progress bar for usage, or will it enforce hard lockouts without warning when large context windows are processed near the limit threshold?

Answer Coming Soon.

We have been receiving requests lately from faculty to purchase a variety of subscription-based services and/or AI platforms. To obtain OIT approval for any of these purchases, should we submit a Data and Technology Review form? Additionally, do you have a decision tree or something that would help us determine which subscriptions and/or AI are allowable and which are not?

The biggest thing to keep in mind is that our security focus is usually less about the specific AI tool itself and more about protecting sensitive data. We have contractual and audit obligations to protect certain types of information (FERPA, PII, etc.), and we want to ensure University of Tennessee data isn't sitting in the history of third-party platforms over which we don't have control. 

Some General Guidance

  • Data Sensitivity: If the data you are using is not sensitive, there are generally very few restrictions on which tools you can use.
  • De-identification: If data is properly de-identified, it opens up more possibilities for tool usage.
  • Meeting Assistants: We have actually blocked many third-party AI meeting bots. Instead, we direct everyone toward Zoom AI Companion and Teams Premium for summarizing meetings. Since people talk about all sorts of things in meetings, this keeps that information within our "closed ecosystem" where it’s safer.
  • Conversation History: It's important to think about where your chat history lives. If it’s with a third party, that party owns that data. If you use our internal tools, it stays on our protected servers.

Recommended Internal Tools

We’ve worked hard to provide access to chat tools that keep our data secure within protected environments. I'd definitely encourage faculty to explore:

  • UT Verse
  • UT AI Hub
  • Microsoft Copilot

To help you and the faculty navigate the thousands of tools out there, we’ve recently updated our Artificial Intelligence guidance page. It covers these principles in much more detail: https://oit.utk.edu/ai/

I would like to know whether I can use Cowork within Copilot at this time.

Copilot Cowork is currently in Research Preview through Microsoft’s Frontier program.  At this time, there are no plans to include it in our Copilot offering. We will monitor the situation once it is publicly released.

 It is problematic that a user cannot upload a file or download a file to work on with the AI systems.  

Hello Howard, thank you for confirming. Unfortunately, this is not a feature that is available within AI Hub. Additionally, it is not something that we can add to the application since it is third party supported but we have passed that suggestion before since others have also requested this. However, the good news is that as we are developing UT Verse version 2, we have made note of your request for this possibility as this is an application we are actively developing ourselves. At present, UT Verse can download entire conversations and responses into MS Word, but it's not an exact replica of the original document. Unfortunately this is a very difficult issue to overcome that even Microsoft has not been able to achieve. 

 

Print Article