Faculty, staff, and students are using AI tools as part of our regular work and productivity. This includes applications such as drafting text (job descriptions, emails, social media, etc.), producing summaries of public information, creating formulas and code, etc. While our usage may overlap, different constituencies on campus have unique considerations.

• Staff should be afforded the capacity to integrate AI into the workplace when it enhances innovation in service of the College’s policies, mission, and values.
• The College also affirms faculty members’ academic freedom and authority to determine whether, how, and to what extent AI tools are used in their courses, assignments, research, and syllabi, within the scope of College policies.
• Students have agency to explore the skills and technologies they may encounter in the present and future.

This is permitted use provided it does not violate the guidance below.  All users should consider themselves accountable for the accuracy, security, and accessibility of materials created with the assistance of AI tools.

Data Classification Guidance

For using institutional data with AI, the College suggests a tiered approach for Data Classification.

• Tier 1
  • Classification: Public (Open)
  • Description: Information with no usage restrictions (e.g., course catalogs, public directories).
  • AI Permissions: Fully Accessible.
• Tier 2
  • Classification: Internal (Restricted)
  • Description: Standard university business data (e.g., department memos, non-sensitive internal reports).
  • AI Permissions: Use licensed, vetted College products only (i.e., K-authenticated Microsoft Copilot with Enterprise Data Protection).
  • This also includes FERPA and FERPA-related student information. Personally identifiable student information (PII) and education records may be used with AI tools only within College-approved systems. This includes systems operated under institutional agreements, such as those provided by Microsoft.
  • Use of identifiable data is permitted only when:
    • The AI tool is accessed through an institutionally licensed and approved environment while the user is logged in and authenticated.
    • The vendor qualifies as a school official under Family Educational Rights and Privacy Act (FERPA).
    • The data is used solely for a legitimate educational or administrative purpose.
  • Only the minimum necessary information is included.
• Tier 3
  • Classification: Highly Sensitive (Confidential)
  • Description: Data protected by law or high risk (e.g., SSNs, financial aid records, health data).
  • AI Status: Prohibited for all Generative AI.

Privacy Risks of AI Tools with Institutional Data

Be aware that the use of AI tools can highlight issues with “oversharing” data in systems like Microsoft SharePoint, Teams, OneDrive, etc.  Common causes of oversharing data access include:

• Setting Privacy for files and folders to “Public.”
• Sharing with all campus users by default (“People in Kalamazoo College with the link…”).
• Granting (inheriting) permissions to someone who should not be able to access this data.
• Inappropriate use of the “everyone except external users” group.

If you are uncertain whether your use of data with AI is acceptable, please discuss this with your supervisor who can help you reach a conclusion. You are the key measure in keeping our data secure. Be mindful of what data you upload and with whom you share it.

In keeping with the College’s commitment to equity, we recognize the importance of equipping all students, regardless of background, with the basic technological knowledge to navigate various workplaces. New technologies can play an important role in promoting access to college resources for students with accommodations as well as first generation students. Generative AI has also introduced new inequities based on access to technology, licensing, and cost.

At the same time, the College is mindful of the cultural biases inherent in AI and seeks to promote literacy around these technologies in which students learn to critically think about the limitations of the information they can access.

Kalamazoo College recognizes that the energy and water demands of Generative AI tools provide a new challenge to sustainability and economic equity. The college community should approach the use of AI with an awareness of the costs to the environment and the importance of climate action. The data to support the energy demands of AI use on our footprint is still emerging. The absence of a reliable metric to measure the negative impacts of AI use is a concern of the Climate Action Plan Committee. However, this does not disregard the truth of AI’s innumerable negative effects on the college’s indirect (Scope 3) emissions.   

There should be an ongoing assessment of the technology measuring AI’s environmental impacts and the role of AI in relation to the College’s commitment to environmental stewardship and climate literacy. Members of our community will utilize artificial intelligence tools – this choice should be informed by concerns about environmental impacts.