Acceptable Use of Generative Artificial Intelligence (GenAI) Policy

Introduction

Mid Michigan College, referred to as Mid, is committed to fully complying with applicable laws and to the ethical use of artificial intelligence (AI). This Acceptable Use of Generative Artificial Intelligence (GenAI) Policy, referred to as the Policy governs the responsible adoption and use of GenAI tools and broader artificial intelligence systems at Mid for business efficiencies, operations, and inclusion in Mid’s products and services.

Purpose

To provide Users with clear guidelines for the acceptable use of Mid AI resources.

Generative AI

Generative AI refers to systems capable of creating new content, such as text, images, code, or media, in response to user prompts (e.g., ChatGPT, Gemini, Microsoft Copilot, note-taking assistance). While this Policy focuses primarily on GenAI due to its widespread availability and specific risks, the principles outlined (lawful use, ethical standards, data protection, transparency, and human oversight) apply to all forms of AI where relevant.

Scope

This Policy applies to all Mid faculty, staff, and contracted staff, collectively referred to as Users. Faculty are encouraged to reference the syllabi options in Appendix II when setting expectations for student GenAI use in their courses. Student use of GenAI is guided by this Policy through Course Syllabi expectations. Specifically for the Writing & Reading Center (WRC) expectations, please refer to Appendix III.

Governance

The Technology Council at Mid is responsible for enforcing this Policy.

Definitions

• Artificial Intelligence or AI indicates the use of machine learning technology, software, automation, and algorithms to perform tasks, set rules, or make predictions based on existing datasets and instructions.
• Artificial Intelligence System or AI System signifies software that is developed with one or more of the techniques and approaches listed in Appendix I. It can, for a given set of human-defined objectives, generate outputs such as content, predictions, recommendations, or decisions that influence the environments they interact with.
• Closed AI System means an AI system in which a user’s input is not used to train the underlying model and remains isolated from other users’ data. These systems are generally considered more secure for handling sensitive information.
• Embedded AI Tools are AI tools that are embedded in existing software tools approved and used at Mid, but which do not require separate approval from the Technology Council. Examples: Microsoft Word’s spelling and grammar checker, Microsoft 365 Suite, Google Workspace, Adobe Creative Suite, and Zoom.
• Generative AI (GenAI) is a type of artificial intelligence that can create new content, such as text, images, or music, based on patterns learned from existing data, rather than just analyzing or predicting. Examples: ChatGPT, Gemini, Claude, and Microsoft Copilot note-taking features.
• Instructional Content is defined as any material used directly in the teaching and learning process. It supports student learning outcomes and is typically created or curated by faculty or instructional staff.
• Non-Instructional Content is any material not directly used in classroom instruction but still important to Mid’s operations, communications, or services.
• Non-Public Mid Data is any information that, if disclosed, could violate the privacy of individuals, government regulations, statutes, jeopardize the financial state of Mid, injure its reputation, or reduce its competitive advantage. Examples include but are not limited to PII, FERPA, and HIPAA data.
• Open AI System means an AI system in which inputs provided by users may be used to train or improve the model. Input data is not fully private and may be exposed to or influence outputs for other users.
• Personal Information is any information that can identify, describe, relate to, or be reasonably linked directly or indirectly with a particular person or household. This includes, but is not limited to, data protected under FERPA, HIPAA, or other privacy laws.
• Traditional AI is a type of artificial intelligence that relies on pre-programmed rules and algorithms to perform specific tasks and does not create anything new. Examples: basic rule-based search engines, virtual assistants using only scripted responses (early versions of Siri, Google Assistant, or Alexa), and simple deterministic tools.
• Technology Council (also referred to as the Technology Steering Committee) is an internal Mid committee that provides executive-level governance and oversight of all institutional technology, including artificial intelligence, at Mid. Its responsibilities include setting technology strategy and priorities, reviewing and approving technology initiatives and projects, recommending technology-related policies and procedures, and
  ensuring alignment with the Mid’s mission and strategic plan.
• Users means all Mid faculty, staff, and contracted staff.

Guiding Principles

The intent of this Policy is to provide general guidance on the use of AI at Mid so that Mid can leverage AI as a tool while ensuring it continues to meet legal obligations and acts in an ethical manner. The use of AI should never compromise Mid’s core values or introduce undue risks. The use of AI at Mid should be focused on improving business efficiencies and enhancing Mid’s ability to fulfill its mission.

This Policy is not intended to address every use of AI. Certain departments and functions bear more considerations and potential risks.

For the use of GenAI in the classroom, faculty may choose from the three syllabus options provided in Appendix II, which will be referenced in individual course syllabi. See Prohibited Uses below for situations where AI may not be used. Additionally, refer to High-Risk Use of AI Systems in below for situations in which extreme caution is required.

Embedded AI Tools used in existing, approved Mid software (for example, spelling/grammar checkers in Microsoft Word) do not require additional approval, provided those tools are aligned with previous general business uses. Explore a current list of approved Embedded AI Tools.

When third-party software, services, or contractors are utilized, any AI usage by these parties must be noted and evaluated carefully. Contracted services that utilize AI technology should be considered in a comparable manner to individual AI usage. Consult with Information Technology/Business Services about the inclusion of an AI-specific clause in any vendor or contractor agreements.

The following principles must be followed when considering the use of an AI system at Mid.

• AI is an advancing technology. We will see it integrate into search engines, email systems, websites, and beyond. If Mid does not manage these tools, they should be treated as Open AI systems. Caution should be used when entering information. Example: Mid manages Google Apps, including GMAIL, and we have some control over how AI is applied. Microsoft email, Yahoo email, and other email systems are not managed by Mid and should be treated as Open AI systems.
• Individuals using an AI system must have expertise in the subject matter for which the AI is being used. Artificial intelligence is to be utilized as a tool and is not a substitute for competence. For example, if GenAI is being used for coding, the individual who is creating the code must have some expertise in coding.
• All AI-generated content (writing, datasets, graphs, pictures, etc.) must be thoroughly reviewed by an individual with proficiency in evaluating such content for accuracy as well as general proofing and editing. AI-generated content should be viewed as a starting point, and not the finished product. AI generated content, especially materials for media, promotion, or public consumption (such as flyers, logos, promotional items, or messaging being posted internally or externally) must be developed and/or approved by the Strategic Communications department prior to distribution or publication.
• AI systems are trained on data that may contain inherent bias. Users of these systems are responsible for reviewing any AI produced content for bias and correcting it, as necessary.
• When integrating AI tools into Mid specific software, users must document all AI systems they are utilizing and for what function. Documentation of approved AI tools is not required. This applies if Mid were to create its own AI assistant.
• The use of an AI system must be documented to capture institutional knowledge. For example, if GenAI is used to create code and included in a larger section of code, there must be documentation as to which code section is AI derived and the person who reviewed it. This documentation can be included within the programming code or completed as a separate document and shared with more than just the owner/creator of the documentation.
• Software or services that integrate AI features being considered for purchase must meet any terms of use or contractual limitations. Contractual restrictions or terms of use may confine Mid’s use of an AI system that would otherwise be legally compliant and ethically sound. For example, an AI system’s terms of use may require certain disclaimers in certain situations or may prohibit the use of the AI system to do certain tasks. Users should have all terms of use or AI systems contracts reviewed by the Information Technology/Business Services to ensure compliance with contractual obligations in using an AI system.
• Approval of an AI system does not eliminate the need for other internal approvals required at Mid for the use of technology, such as a security review, privacy review, cost review, spending approval, legal review, human resources review, etc. An AI system should go through the same review and approval process as other software or services at Mid. Reference General AI System Use Standards & Use Approval section below.

Prohibited Uses

Users are prohibited from using AI systems for any of the following activities at any time (unless otherwise approved by the Technology Council and respective department heads).

• Use of AI systems in meetings where Non-Public Mid data or personal information is discussed, including but not limited to confidential or sensitive meetings.
• The use of third-party AI tools, such as meeting recorders, note-taking tools, and writing proofreaders, is only permitted if the tool has been approved and all meeting participants have agreed to its use.
• Using AI systems to identify students, candidates, employees, contractors, or any type of person.
• AI should not be utilized in any manner related to evaluating potential candidates for admission into Mid, internship or apprenticeship programs, or any other Mid program.
• Putting non-public Mid data into an Open AI system.
• Entering confidential information or personal information about any individual into an Open AI system. Includes medical, financial, political affiliation, racial or ethnic origin, religious beliefs, gender, sexual orientation, disability status, or any part of a person’s private life or legal status. Examples include but are not limited to FERPA, HIPAA, etc.
• Using Mid time or resources to generate content using an AI system that would be considered illegal, inappropriate, or harmful to Mid’s brand or reputation. Using AI systems that would be disrespectful to others is prohibited.

Ethical Guidelines

Mid seeks to act in an ethical manner when using AI. Accordingly, there may be uses of AI that are legally permissible, but which do not meet Mid’s ethical standards. Any use of an AI system at Mid should conform to the following ethical guidelines.

• Integrity in Use
  • All Users of AI systems should be forthright and transparent about how AI assisted in completing the work. Even if using an AI system approved by the Technology Council, the supervisor of the department requesting a task for which you are using an AI system should be aware of your use of the AI system. Do not misrepresent AI generated work as solely yours. For example, when writing a performance evaluation, the supervisor and/or HR representative should give permission for AI system assistance.
  • Research Use
    • When using data related to students or Users, the data needs to be reviewed by
      Institutional Research to ensure that the data is anonymized and is being used for educational purposes.
  • Unauthorized Use
    • Do not use Mid time or resources to generate content using an AI system for
      personal use without prior approval of the appropriate department leader.

High-Risk Use of AI Systems

There are certain uses of AI systems that are higher risk than others.

• Personal Data in AI Systems
  • AI should be used with extreme caution when inputting any personal data of an individual into a Closed AI system (it is prohibited to put any personal information into an Open AI system).
• Personnel Decisions
  • AI should be handled with caution for any use related to making decisions on promotions, retention, or similar personnel decisions. Extreme caution should be utilized to ensure that biases (including biases found in existing data sets and any forms of implicit bias) are avoided.
• Assessment of Students
  • Any assessment of students in educational or vocational training institutions is considered high risk. Accordingly, extreme caution should be taken before using any AI system intended to assess or evaluate any student participating in a course, taking an exam, or any other evaluation or assessment.

General AI System Use Standards & Use Approval

Use of any AI system(s) for Mid business should follow the guidelines of this policy, and if it is a product being considered for purchase or software as a service, the Technology Council must approve it to ensure that such AI systems meet the following principles.

• Lawful
  • The use of AI systems must comply with all applicable laws and regulations, as well as any contractual obligations, limitations, or restrictions.
• Ethical
  • The use of AI systems must adhere to ethical principles, be impartial, and avoid bias or disclose found implicit bias.
• Transparent
  • There must be clear objectives for the use of an AI system and documented oversight of such use. This should be recorded and captured for institutional knowledge. Disclosures of the use of GenAI in any AI assisted content generation must be made when required by law or contract or when required by Mid.
• Necessary
  • The use of AI systems must be for a valid business purpose to improve Mid’s business efficiencies and support Mid’s mission. The use of AI is not a substitute for human critical thinking or expertise. AI systems should not require Mid to incur an unnecessary expense without a valid benefit.

Prior to submitting a request to the Technology Council for the use of an AI system, a requester should first obtain the approval of their supervisor. In addition, in evaluating whether to make a request prior to submitting it, the requester should ensure that the AI system use (if approved) would conform with the guidelines in this Policy. Requests for the use of an AI system should follow the steps below.

• Request supervisor approval to proceed with the request.
• Provide details about the solution to Information Technology for review and feedback.
• Present the recommendation to the Technology Council

Explore a List of Approved AI Systems

Training

All Users who interact with AI systems are expected to be familiar with this Policy. Mid provides training and resources to support the responsible use of Generative AI. Users should complete any required AI training, which may cover responsible use, data privacy, bias mitigation, accuracy of AI outputs, and proper disclosure practices. Additional resources and guidance will be available through Information Technology and Human Resources.

Reporting Non-Compliance 

Mid Users that are aware of any conduct that may violate this Policy have a responsibility to report it. Individuals are encouraged to make reports through normal reporting relationships beginning with their supervisor. The Technology Council, Legal Counsel, Human Resources, or other appropriate parties will investigate all reports of suspected misconduct or non-compliance. Unless they act in bad faith, Mid Users will not be subject to reprisals for reporting potential violations. After an investigation concludes, if Mid determines that a User has failed to comply with this Policy, the User will be subject to disciplinary action, per the respective College policies.

This policy shall be reviewed and updated as needed to reflect changes in AI technology, ethical guidelines, and legal requirements.