Home
Faculty & Staff Services
AI Literacy Initiative

FAQ

Getting Started

Generative AI tools (e.g., Copilot, Claude, ChatGPT, Grok, etc.) create text, summaries, images, or code in response to a prompt (a request you make). They work by predicting a likely response from patterns derived from huge amounts of training data. That makes them fast and flexible, but it also means they can sound confident while being wrong.

Remember: AI works best as a collaborative tool, and you bring what it can’t—context, judgment, and accountability.

Am I allowed to use AI in my work?

Yes, with some limitations in accordance with OIS policy. Generative AI tools are not a private workspace. Without appropriate data protections, what you share with ChatGPT, Claude, Gemini, or other tools can be stored by the vendor and used to train future models.

University-approved tools like Copilot and Claude come with contracts that limit how your data can be used while you are logged in with your university credentials.

This table illustrates which AI tools may be used with which types of university data. Note that certain types of data (FERPA, HIPAA, etc.) may have different rules.


Category	Can be used with
Category I — Confidential	No AI tools
Category II — Controlled	AI tools with enterprise data protection enabled (Copilot, Claude)
Category III — Published	Any AI tool that is not prohibited by Texas Cyber Command

View the full AI tools and data table

For questions about the OIS Standard for Acceptable Use of Generative AI Tools, please email InformationSecurity@utsa.edu

Which AI tools can I use?

The tool you can use depends on the type of data you are working with.

Public data (Category III (opens in a new window))If you are working with publicly available data, you can use any type of AI tool that is not prohibited by Texas Cyber Command.
Confidential and Internal data (Category I and II (opens in a new window))UT San Antonio has enterprise agreements for using confidential (category II) data with Microsoft Copilot (and Copilot Premium), and Claude. Internal data (category I) should not be used with generative AI tools.

All UT San Antonio students, faculty and staff have access to Copilot Chat as part of their Office 365 productivity tools. With purchase, faculty and staff can access Copilot Premium or Claude. To request access to Copilot Premium or Claude, fill out this request form.

How do I get access to AI tools?

UT San Antonio provides all students, faculty, and staff with Microsoft Copilot Chat at no additional cost.
Get started with Copilot Chat»
With purchase, faculty and staff can access Microsoft Copilot for Office 365 (Premium), which comes with additional features and file integration.
Get started with Copilot Premium»
In addition, with purchase, faculty and staff can access Claude.
Get started with Claude»

To request purchase of a license for Copilot Premium or Claude:

Fill out the purchase form » (opens in a new window)

Is there required AI training?

In 2026, the state of Texas launched new mandatory IT security training, including a statewide Artificial Intelligence (AI) Awareness course now required by Texas House Bill 3512 for AI literacy for state employees. This training must be completed by all faculty, staff and student employees at UT San Antonio.

Take the mandatory AI Awareness Training

Data, Privacy, and Security

What information can I put in an AI tool?

The information you can put in an AI tool depends on two factors: the sensitivity of the data and the tool you'd like to use.

Public dataIf you are working with publicly available data, you can use any type of AI tool that is not prohibited by Texas Cyber Command.
Internal or Confidential dataIf you'd like to use AI with internal data, you must use a tool for which UT San Antonio has a data protection agreement in place. Confidential data should not be placed in any AI tools.

Per OIS Standard 09—Data Classification (opens in a new window):

All university information resources must be classified by Data Owners into one of three categories—Category I (Confidential), Category II (Internal), or Category III (Public).

Category I (Confidential): Data protected by federal/state law or contract (e.g., FERPA, HIPAA, SSN, PCI, research IP).
Category II (Internal): Non-public operational data; unauthorized disclosure causes institutional harm.
Category III (Public): Intended for public release; no access restrictions.

Can I use AI with student records (FERPA) or health information (HIPAA)?

Some AI tools available at UT San Antonio, like Microsoft Copilot, come with enterprise data protection. This means your prompts and files aren’t used to train the underlying models, and they stay within our institutional agreements. That is an important safeguard, but it isn’t the same as approval to use every kind of data in these tools.

We’re still confirming official guidance on which data classifications are permitted. Until we have it, best practice is to keep sensitive and regulated data out of AI tools. This includes:

Social Security numbers and other government ID numbers
Student records protected by FERPA (grades, enrollment status, ID numbers tied to names)
Health information, especially relevant as we integrate with UT Health San Antonio
Financial account information, payroll data, or anything covered by contractual or legal confidentiality

If you need AI help with a task that involves this kind of data, you can often still use it safely by:

Replacing real values with placeholders, such as “[Student Name]” or “XXX-XX-XXXX”
Describing the structure or problem instead of pasting the records themselves
Using a small, made-up sample to work out the approach, then applying it yourself

We’ll share updated guidance once it’s confirmed. Until then, if you’re unsure, leave it out.

Does the AI tool keep or learn from what I type?

Free consumer tools may store your conversations and use them to improve their models. University-licensed tools are covered by contracts that restrict this, called enterprise data protection agreements. If you are working with any data that is not publicly available, be sure you are signed in to one of these tools with your university credentials: Copilot Chat, Copilot Premium, or Claude.

Copilot Chat is available at no additional cost for all students, faculty, and staff. Copilot Premium and Claude are available by department purchase. Request paid access to these tools here.

How can I tell whether I'm using the protected version of Copilot?

Check that you're signed in with your UT San Antonio account. In Copilot, Microsoft shows a shield icon when enterprise data protection is on. If you don't see your university account or the shield, you may be in the personal version.

green Copilot shield indicating data protections are on

What if I accidentally put something sensitive into an AI tool?

It depends on what you shared. Some situations need quick action, some need a report, and some just need cleanup.

If you shared a password, API key, or other credential
Change it right away. Don’t wait to report first. A credential that has left your control should be treated as compromised, and changing it is the one fix that fully works.
If you shared regulated data, such as patient health information, student records, Social Security numbers, or financial account information
Report it to InformationSecurity@utsa.edu as soon as possible. Include which tool you used, roughly when, and what kind of information was involved.

Laws such as HIPAA, FERPA, and Texas data-breach rules require UT San Antonio to assess certain disclosures within set deadlines. Reporting promptly protects you, the people whose data it was, and the university.
For anything else, such as internal drafts, meeting notes, or non-public but unregulated information
Delete the conversation from the tool’s history. You usually don’t need to file a report. If you’re unsure whether something counts as regulated, email InformationSecurity@utsa.edu.

Can I accept the terms of service for a new AI tool myself?

Usually not. Clicking “I accept” on an AI tool’s terms of service creates a contract, and only employees with a written delegation of authority from the president can sign contracts for UT San Antonio.

For more information, see these resources:

Clickwrap Agreement Financial Guidelines (opens in a new window)
HOP 11.07 Cloud Computing (opens in a new window)

Using AI Responsibly

Can I trust what AI tells me?

Not without checking. AI tools can invent facts, sources, policies, and numbers, and present them in the same confident tone as accurate information. Verify anything that matters, especially names, dates, figures, citations, and anything about university policy.

Do I need to tell people when I used AI?

A good rule is to be open about AI use whenever it would matter to the person receiving the work. For example, say so when you used AI to draft a public-facing document, summarize someone's words, or produce analysis others will rely on.

Can AI make decisions about people?

No. Don't let AI make or control decisions about hiring, performance, discipline, admissions, financial aid, grades, or similar matters that affect someone's opportunities. Texas law and OIS policy require meaningful human review of these "consequential decisions."

Can AI be biased?

Yes. AI tools learn from human-written data, so they can repeat stereotypes or leave out perspectives. Read outputs critically, especially anything about people, groups, or communities.

What about copyright?

Be careful with both directions. Don't upload copyrighted material you don't have rights to use, such as purchased reports or other people's unpublished work. And review AI-generated text and images before publishing them, since you, not the tool, would likely be responsible for any infringement.

Getting Help

I'm new to AI. Where should I start?

Pick one small, low-risk task you do often, like drafting a routine email or summarizing a public report, and try it with Copilot Chat. Then try some of these suggested prompts or drop in to our open office hours. You don't need technical skills to start.

How do I write a good prompt?

Give the tool the context a new coworker would need. Say who the audience is, what you want produced, how long it should be, and what tone to use. If the first answer isn't right, tell it what to change rather than starting over. Iterate as many times as it takes to get what you're looking for.

Where can I get training?

UT San Antonio is pleased to offer many resources for learning about AI:

From Enterprise Technology Services

Suggested Prompts (Prompting Tips and Tricks)
Virtual/in-person live training sessions (opens in a new window)

From Academic Innovation

Using Generative AI in Teaching, Learning and Research Guidelines by and for UT San Antonio Faculty
AI Teaching and Learning Resources

Available at no cost to UT San Antonio faculty and staff:

Artificial Intelligence Training (Coursera) (opens in a new window)
Artificial Intelligence Courses (LinkedIn Learning) (opens in a new window)

Where can I ask questions?

If you have questions about AI use cases, feel free to attend any of our drop-in Office Hours sessions (Tuesdays at 11 a.m. CT), ask in our AI Literacy Hub Teams site, or email techcafe@utsa.edu.

Faculty are also welcome to join the AI Peer Learning Network

Join Office Hours
Join the Teams Site

For questions about data security or acceptable use of AI tools, please email InformationSecurity@utsa.edu.

FAQ

On this page