AITS AI Guidelines for Safe Use

These guidelines support the responsible, secure, and compliant use of artificial intelligence (AI) tools by AITS and System Office staff working with University data. Approved AI services have been vetted for security, legal, and data privacy considerations and are intended to help employees use AI effectively without introducing unintended risk. Typical use cases include (but are not limited to): data analysis, automation, research support, code and script generation, chatbots, and productivity assistance.

Scope: These guidelines apply to AITS and System Office employees, contractors, and approved external participants who use AI tools with University data.

Use Approved Tools Only

Use only AI platforms and services that have been approved for handling University data.
Do not use unapproved AI tools for University work.
Personal versions of tools (e.g. ChatGPT, Gemini, Claude, Copilot, Perplexity) should be avoided with University data. Unchecking e.g. "Use my data to train the model" or "Help improve the model for everyone" is generally advised for personal tools.
Ensure the service has completed required security, legal, and data privacy reviews.
Follow any usage instructions, constraints, or limitations provided by AITS, Information Security, or Data Governance teams.

Understand the Data You’re Using

Only use AI tools with data you're allowed to share. If you're unsure, please contact AITS AI Requests aitsairequests@uillinois.edu.

✅ Public

Information intended for public release. No adverse impact if disclosed.

Generally appropriate for AI

⚠️ Internal

Non-public university data intended for internal use. Potential operational impact if disclose.

Use with caution

⚠️ Sensitive

Role-restricted or sensitive university data. Serious privacy or operational risk if improperly shared.

Approved tools only

🚫 High Risk

Regulated or high-risk data (e.g. SSNs, credit cards, protected health information). Do not enter into AI tools unless that tool has special approval or you've received an exception.

Do not use with AI

Know how your data is classified: high-risk, sensitive, internal, or public.
Examples of sensitive or high-risk data include PII, PHI, FERPA-protected records, SSNs, financial data, and proprietary information.
Be aware of data volume (e.g. tens of thousands of records).
Approval of a tool does not imply unrestricted use of sensitive or high-risk data.

The detailed list of approved tools and their restrictions is available in the AITS Knowledge Base (Authentication Required).

Requesting Review of Unapproved Use Cases or Tools

If a desired use case, AI tool, or service is not listed above:
1. Complete the AITS AI Risk Assessment Form. AITS will assist with or guide completion of the assessment as needed.
2. Complete the Chief Digital Risk Officer (CDRO) AI Discovery Form.
The AITS AI team may assist with implementation where appropriate.
Enterprise Architecture implications will be reviewed as part of the process.

Prepare and Submit Data Thoughtfully

Minimize data shared with AI tools: submit only what is necessary for the task.
For testing or experimentation, use anonymized, synthetic, or sample data instead of real records.

Review and Validate AI Outputs

AI-generated outputs must be reviewed by a human before use in decision-making.
Do not rely solely on AI outputs for critical, operational, or compliance-related decisions without appropriate human review.
Validate summaries, recommendations, code, and predictions.
Document how AI outputs were used, where appropriate.

Protect Outputs and Share Responsibly

Do not share AI-generated outputs that include sensitive data via public channels (e.g., non-University email, chat, personal accounts) unless explicitly allowed.
Disclose AI use where appropriate, especially in reports, official communications, or externally shared materials.
Maintain records of AI tools used for sensitive projects.
Follow University guidance on ethical and responsible use of generative AI.
When in doubt, consult AITS AI Requests (aitsairequests@uillinois.edu).

Training

Some sources of general AI training are provided below. Please be sure to check this page regularly for training updates.
AI Professional Development Series (for faculty and staff via Illinois Human Resources)
LinkedIn Learning Paths (free with Illinois login)
Introduction to Generative AI (via University Library)

Human Resources Guidelines

System Office Human Resources provides guidance on the use of artificial intelligence (AI) in employment and hiring processes for the University of Illinois System, ensuring compliance with state law and university policies.
For more information see Hiring and Employment

Report Issues and Concerns

Please direct questions, unexpected behavior, suspected data exposure, or security concerns to AITS AI Requests (aitsairequests@uillinois.edu).
Notifications are reviewed by the AITS AI team and AITS Enterprise Systems Assurance.