Guide to Artificial Intelligence (AI)

Summary

This document is intended to help provide guidance on how best to safely use AI with your work. It is not intended to define acceptable use of AI in any academic endeavor.

Body

Purpose

This document is intended to help provide guidance on how best to use AI safely with your work. It is not intended to define acceptable use of AI in any academic endeavor.

Definitions

Artificial Intelligence or AI

Any form of AI including ChatGPT, Gemini, Copilot or other systems that leverage prompts or other data input to generate text, image, or other content.

Prompt

Any question or task presented to an AI including any data provided to the AI.

Sensitive Data

Any information classified as sensitive under UW System policies or private to an individual or could put the University at risk if it were shared publicly. Examples would include names, grades, addresses, titles, ranks, etc.

Personally Identifiable Information (PII)

Personally Identifiable Information is information that can uniquely identify a person. Generally, it includes their name along with one or more of the following pieces of data:

  • Social Security Number (SSN)
  • Employee or Student ID number
  • Driver’s License Number
  • State issued ID card number
  • Bank or other financial account information
  • Medical or health insurance information

Incident

An Incident is any event where sensitive, personal, or other restricted data was shared with any AI, intentionally or accidentally.

Guidance

Confidentiality

All private and sensitive data, including personal identifiers, academic records, research data, and health information, must be handled with the highest degree of confidentiality and should not be used with any current AI system or service without prior approval from campus leadership and ITS.

Assume any data shared is being shared with the public.

Data Minimization

When constructing an AI prompt, use and share only the data necessary for the specific purpose.

Bias

AI systems are not immune from bias. Be certain to review all results for both accuracy as well as unintended biases generated by the AI.

Awareness

Ensure your staff and students are aware of this policy along with the risks and cautions created by using AI systems.

Incident Reporting

Any suspected breaches of data privacy or security, intentional or accidental, should be reported to ITS via the Help Desk immediately.

Legal and Ethical Use

Any use of any AI system must comply with all legal requirements and UW System or campus policies. Additionally, any illegal or unethical uses or prompts are strictly prohibited.

Dos and Don’ts

The following are some “dos” and “don’ts” to help guide usage of AI on campus:

Do:

  • Remove any personally identifiable data
  • Assume all data shared is being shared publicly
  • Use reputable and well-known AI services
  • Follow all regulations and policies
  • Train yourself, your staff, and your students on safe and acceptable use of AI

Don’t:

  • Use illegal or unethical prompts
  • Share internal documents or data
  • Assume others are aware of the risks of using AI
  • Neglect to report any security incidents

UW-System Resource

Generative AI | Information Technology & Security

 

Details

Details

Article ID: 15690
Created
Tue 10/29/24 3:35 PM
Modified
Tue 10/29/24 4:18 PM

Related Articles

Related Articles (2)

Instructions for logging into Microsoft Copilot using your UWPLATT account.
Features, resources, and security tips related to Zoom AI Companion at UW-Platteville.