5 Things People Should Do When Considering AI
- Think about how you want to use it.
- Teaching
- Research
- Learning outcomes
- Other (e.g., process improvements, etc.)
- Decide how you will procure it.
- Develop a new resource using currently available U-M AI tools?
- Buy it, with Ross IT assisting with the procurement.
- Know the strengths and weaknesses of AI.
- Define the problem you are trying to solve. – AI is a tool, not a goal.
- AI excels at pattern recognition, parsing information, and handling large datasets.
- It is less reliable with nuanced judgment, up-to-date real-world context, or solving ill-defined problems.
- AI systems rely on accurate, relevant, and clean data.
- Plan for maintaining data privacy, availability, quality, and academic integrity.
- Only use university-approved AI platforms for sensitive content.
- Understand data ownership and compliance requirements (GDPR, HIPAA, FERPA)
- If using external AI products, ensure you understand what university data you can share and how it needs to be protected. Any external product must adhere to the following policies.
- Information Security (601.27)
- Vulnerability Management (DS-21) (Cloud and Third Party Vendor Applications)
- Sensitive Data Guide to IT Services
- Library-licensed journal papers, etc., cannot be entered into third-party AI research tools like Elicit (publishers forbid that in our licensing terms). Only use UM AI tools for that.
- Ensure Accessibility and Equity
- Consider whether all your students and collaborators can access the AI tools you require or assign.
- Learn about AI’s biases and how they may influence responses.
Last Updated on August 18, 2025
