Getting Started With Artificial Intelligence (AI) at University of St. Thomas

Summary

Information about Artificial Intelligence use at University of St. Thomas.

Body

Guidance may change: AI tools, contracts, security controls, pilots, licensing, and approved university uses continue to evolve. This article provides general direction and does not replace university policy, academic requirements, research protocols, records-retention obligations, contractual restrictions, or instructions from an authorized university data owner or data steward.

Issue or Question

What should members of the University of St. Thomas community know before selecting and using an artificial intelligence tool?

Artificial intelligence (AI) can help members of the St. Thomas community generate ideas, organize information, revise content, analyze approved data, improve accessibility, support teaching and learning, and streamline administrative work.

This article provides an overview of responsible AI use, university data classifications, available AI tools, academic considerations, and resources for additional guidance.

AI-generated content can be inaccurate, incomplete, biased, outdated, or presented without sufficient context. The person using an AI tool remains responsible for protecting university information and reviewing the final result.

Environment

This information applies to:

  • University of St. Thomas faculty, staff, and students
  • Student employees, contractors, volunteers, and other individuals conducting university business
  • AI tools used for teaching, learning, research, or administrative work
  • University-provided, university-managed, individual, and consumer AI tools
  • University data entered into prompts, chats, files, assistants, agents, integrations, or connected applications

Resolution

Before You Begin

Important: Before entering information into an AI tool, identify the university data classification and confirm that the specific tool is authorized to process that classification. Do not assume that a paid account, enterprise label, privacy setting, or university sign-in automatically makes a tool appropriate for sensitive university information.

AI at St. Thomas

St. Thomas approaches AI as both an opportunity and a responsibility. University AI initiatives seek to promote AI literacy, thoughtful innovation, academic excellence, ethical decision-making, and responsible technology use.

  • The Institute for AI for the Common Good coordinates AI-related activities, supports research and development, encourages dialogue, and helps align AI use with university values and priorities.
  • The Center for Applied Artificial Intelligence, located within Graduate Programs in Software in the School of Engineering, supports applied research and collaboration involving AI, machine learning, data management, data visualization, and cloud infrastructure.
  • The Strategic Transformation in Education, Learning, and Research (STELAR) team supports instructional design, academic technology, emerging technology, research computing, and the thoughtful use of technology in teaching and learning.

Responsible AI Use Principles

  1. Protect university data. Determine the data classification before entering information into a prompt, uploading a file, creating an assistant, or connecting an AI tool to another application.
  2. Use the appropriate tool. Select a university-approved environment that is authorized for the information involved.
  3. Maintain human oversight. Review AI output for accuracy, completeness, bias, privacy, security, accessibility, copyright, and suitability for the intended audience.
  4. Use the minimum information necessary. Remove identifying details, summarize information, or use approved sample data whenever the task does not require the complete record.
  5. Be transparent when appropriate. Disclose meaningful AI assistance when required by an instructor, department, research protocol, publisher, funding organization, professional standard, or other applicable process.
  6. Follow academic expectations. Permission to use AI in one course, assignment, or activity does not automatically extend to another.
  7. Retain human accountability. The person using AI remains responsible for the final content, communication, recommendation, or decision.

Select an AI Guidance Topic

Select a section below to show or hide its information.

Classify the most sensitive information that the AI tool could receive. Consider prompt text, uploaded documents, images, recordings, transcripts, email, meeting content, system exports, connected application data, and information retrieved by an AI assistant or agent.

Green Data: Public

Green Data is information intended or approved for public access.

Examples include:

  • Published press releases
  • Public campus maps
  • Approved marketing materials
  • Published athletic results
  • Public university webpages
  • Publicly available course or program information

General direction: Green Data may be used only in AI tools and use cases that otherwise comply with academic, copyright, accessibility, records-management, and university requirements.

Yellow Data: Internal

Yellow Data is nonpublic information intended for use within the university.

Examples include:

  • Internal procedures and planning documents
  • General system information
  • Internal memos
  • Routine business email
  • Unpublished research that is not otherwise restricted

General direction: Use Yellow Data only in university-approved tools that are authorized for Yellow Data. Do not enter Yellow Data into an individual or consumer AI account.

Orange Data: Confidential

Orange Data is sensitive university information requiring additional protection and limited access.

Examples include:

  • Family Educational Rights and Privacy Act (FERPA)-protected student data
  • Employee records
  • Donor giving history
  • University contracts
  • Sensitive operational or financial information
  • Confidential research information

General direction: Use Orange Data only in a specifically authorized university-managed AI environment. Provide only the minimum information necessary and follow restrictions established by the applicable data owner, data steward, policy, contract, research protocol, or business process.

Red Data: Highly Restricted

Red Data is the university's most sensitive information. Unauthorized access or disclosure could create significant legal, financial, safety, privacy, or institutional consequences.

Examples include:

  • Social Security numbers
  • Payment card information
  • Protected health information
  • Individually identifiable human-subject research data
  • Passwords, security keys, authentication codes, and other authentication secrets
  • Other information designated as highly restricted by university policy

Caution: Do not place Red Data into an AI tool unless the university has explicitly authorized the specific tool, account, configuration, and use case. A general product name or data-use designation does not provide blanket permission to process all Red Data.

Mixed classifications: When information contains more than one classification, handle the complete collection according to the highest classification present. For example, a document containing both Green Data and Orange Data must be treated as Orange Data.

The following matrix summarizes the university AI tool and data-use information available when this article was prepared. Tool availability, licensing, pilots, security controls, and approved data uses may change.

AI tools and university data-use overview
Tool or Account Type Green Yellow Orange Red
Individual or consumer AI tools, including personal ChatGPT, Claude, Gemini, or Copilot accounts Yes No No No
Aquinas AI pilot Yes Yes No No
BoodleBox university workspace Yes Yes Yes No
University-provided ChatGPT Business Yes Yes Yes No
University-managed Claude Team Yes Yes Yes No
University-approved Microsoft Copilot experience Yes Yes Yes Conditional

Caution: Red Data use in Microsoft Copilot depends on the specific Copilot product, university account, configuration, connected services, and business process. Contact ITS Information Security before proceeding with a Red Data use case.

Consumer AI Tools

Individual and consumer AI accounts should be treated as public-data environments. Do not enter internal, confidential, or highly restricted university information into a personal ChatGPT, Claude, Gemini, Copilot, or similar account.

Aquinas AI Pilot

Aquinas AI is an internal university AI pilot available to faculty, staff, and students with active St. Thomas accounts. Published pilot information states that interaction data is retained for 90 days and then deleted in accordance with university data-management practices.

Interaction data is treated as private under ordinary circumstances but may be recoverable and subject to review in connection with an authorized university investigation, legal proceeding, or other authorized inquiry. Aquinas AI data is stored in the University of St. Thomas Amazon Web Services environment.

Do not use Aquinas AI for sensitive or confidential information beyond what is expressly permitted by university guidance and the Data Security Classification Policy.

BoodleBox

BoodleBox is a collaborative AI platform designed for higher education. The St. Thomas environment can provide access to multiple AI models, collaborative workspaces, custom assistants, document-based knowledge features, and responsible-use resources.

Access is subject to university licensing and approval. Use the university workspace assigned to the St. Thomas account rather than a separate personal workspace.

ChatGPT Business and Claude Team

When university access is provided, use only the university-managed business or team environment for university information. A personal account for the same product does not have the same university authorization.

Microsoft Copilot

Use only a university-approved Copilot experience while signed in with the appropriate St. Thomas account. Data protection can vary among Copilot products, features, agents, plugins, connections, and account types.

  1. Identify the activity. Determine whether the activity supports university business, teaching, learning, or research.
  2. Identify all information the AI tool could receive. Consider prompts, files, images, recordings, transcripts, email, chat, meeting content, system exports, personal details, research data, and connected applications.
  3. Determine the data classification. Classify the most sensitive information involved. When several classifications are present, use the highest classification.
  4. Confirm that the tool is authorized. Verify the specific product, university account, workspace, configuration, and intended use.
  5. Identify additional requirements. Determine whether the information is subject to FERPA, research requirements, employee privacy, health-information protections, payment-card rules, contracts, copyright, donor restrictions, export controls, or other requirements.
  6. Reduce the information supplied. Remove direct identifiers, use approved sample data, summarize instead of uploading complete records, and provide only what the task requires.
  7. Review the output. Verify important facts, sources, citations, quotations, links, calculations, names, dates, recommendations, and technical instructions.
  8. Document or disclose AI use when required. Follow applicable instructions from the course, instructor, department, research protocol, publisher, sponsor, or professional standard.

Commonly Suitable Uses

The following activities may be suitable when the tool is authorized for the information involved:

  • Brainstorming topics, questions, or alternatives
  • Creating outlines and initial drafts
  • Revising tone, organization, grammar, and readability
  • Summarizing approved material
  • Generating practice questions or study prompts
  • Creating formulas or code prototypes
  • Developing meeting agendas or prompt templates
  • Identifying themes in approved data
  • Translating or simplifying public content
  • Creating accessibility-focused alternatives followed by qualified human review
Uses Requiring Additional Review

Consult the appropriate university resource before using AI for:

  • Admission, employment, financial aid, grading, discipline, promotion, or access-to-service decisions
  • Legal, compliance, health, safety, or financial determinations
  • Student, employee, donor, or alumni profiling
  • Human-subject research
  • Automated communications issued on behalf of the university
  • Public-facing institutional statements
  • Procurement or implementation of a new AI service
  • Integration with university systems or institutional data
  • Creation of a department-wide assistant, agent, or knowledge repository
  • Any use that could materially affect an individual
Uses to Avoid
  • Fabricating citations, evidence, records, research results, or sources
  • Misrepresenting AI-generated work as independently authored when disclosure is required
  • Entering university information into an unauthorized tool
  • Sharing passwords, security keys, authentication codes, or account-recovery information
  • Making consequential decisions without qualified human review
  • Circumventing academic, workplace, research, or professional expectations
  • Uploading information solely because the user has access to it
  • Reproducing copyrighted or licensed material beyond permitted use

Students
  • Review the course syllabus and assignment instructions before using AI.
  • Ask the instructor when AI expectations are unclear.
  • Do not assume that permission to use AI in one course or assignment applies to another.
  • Retain notes showing how AI was used when disclosure may be required.
  • Verify generated information, citations, calculations, and sources.
  • Do not upload another person's private information, coursework, or intellectual property without appropriate permission.
Faculty
  • Clearly communicate whether AI use is prohibited, limited, permitted with acknowledgment, or expected.
  • Design assignments around the intended learning outcomes.
  • Protect student work and FERPA-protected information.
  • Do not rely solely on automated AI-detection tools when reviewing student work.
  • Consult STELAR for instructional-design, academic-technology, and emerging-technology guidance.
Staff and Administrators
  • Treat AI output as a draft unless an approved university process states otherwise.
  • Use university-managed tools for university business.
  • Consult the applicable data owner or data steward before using departmental datasets.
  • Do not use AI to make final consequential decisions about an individual.
  • Follow technology intake, purchasing, security, accessibility, records-retention, and contractual requirements.
Researchers
  • Follow applicable research protocols, data-use agreements, sponsor requirements, consent language, and publication standards.
  • Do not enter identifiable human-subject data into an AI tool unless the environment and specific use have been explicitly approved.
  • Document material AI contributions to research design, coding, data analysis, writing, visualization, or other research activities.
  • Verify whether AI-generated output may be used in the intended project or publication.
  • Consult the principal investigator, Institutional Review Board, research office, sponsor, or applicable data steward when requirements are unclear.

Accuracy

AI systems can produce confident but incorrect responses. Independently verify names, titles, dates, deadlines, policy language, statistics, calculations, technical instructions, citations, quotations, links, and legal or compliance claims.

Bias and Fairness

Review generated material for assumptions, stereotypes, missing perspectives, or patterns that could unfairly affect individuals or groups. Use additional caution when AI output relates to evaluation, selection, discipline, advising, employment, assessment, or allocating resources.

Copyright and Intellectual Property
  • Do not assume that AI-generated material is original, unrestricted, or appropriate for publication.
  • Avoid requesting substantial reproduction of protected works.
  • Check licensing requirements for source material.
  • Attribute sources when appropriate.
  • Confirm publication and commercial-use rights.
  • Protect unpublished university work and intellectual property.
Digital Accessibility

AI can assist with headings, summaries, plain-language revisions, alternative text, captions, and transcripts, but automated output requires human review. Confirm that final content accurately represents the source and meets applicable university digital-accessibility requirements.

AI Strategy and Universitywide Coordination

Consult the Institute for AI for the Common Good for information about university AI initiatives, working groups, ethical dialogue, research, and institutional coordination.

Teaching and Learning

Contact STELAR for course design, instructional technology, emerging technology, research computing, Canvas, and AI-supported teaching guidance.

Tool Access and Technical Support

Use the OneStThomas service catalog or contact the Tech Desk for university AI account access, licensing, sign-in problems, and troubleshooting for university-provided AI services.

Data Security and Privacy

Contact ITS Information Security when a proposed use involves confidential or highly restricted information, system integrations, uncertainty about data classification, a new AI product, or a possible security or privacy incident.

Research

Consult the appropriate research office, Institutional Review Board, principal investigator, data steward, or sponsor when AI will be used with research data or as part of a research methodology.

Quick Reference Checklist

Before using AI for St. Thomas work, confirm the following:

  • I am using a university-approved AI tool.
  • I am signed in with the correct university-managed account and workspace.
  • I have identified the classification of all information involved.
  • The tool is authorized for that classification and use case.
  • I am providing only the minimum information necessary.
  • I have permission to use or upload the information.
  • I have considered applicable course, department, research, contractual, legal, or data-owner requirements.
  • I will independently review and verify the output.
  • I will review the result for bias, privacy, security, copyright, and accessibility concerns.
  • I will disclose the use of AI when required or appropriate.
  • I can explain the AI use and resulting decision to affected individuals or reviewers.

If any required answer is no or uncertain: Pause the activity and consult the appropriate university resource before entering information into the AI tool.

Troubleshooting

If you cannot determine whether an AI tool or use case is appropriate:

  • Confirm that you are signed in with the correct university-managed account and workspace.
  • Confirm the classification of the most sensitive information the tool could receive.
  • Remove identifying details and information that is not required for the task.
  • Review course, department, research, contract, and data-owner requirements.
  • Do not enter the information until the tool and use case have been confirmed as authorized.
  • Contact the Tech Desk or ITS Information Security when the appropriate tool or data classification remains unclear.

Additional Resources

Related Information

Use the following university resources for additional information and assistance:

Guidance may change: AI tools, contracts, security controls, pilots, licensing, and approved university uses continue to evolve. This article provides general direction and does not replace university policy, academic requirements, research protocols, records-retention obligations, contractual restrictions, or instructions from an authorized university data owner or data steward.

Details

Details

Article ID: 170872
Created
Wed 9/30/26 2:46 PM
Modified
Thu 10/22/26 1:08 PM