ApplyVisitGive
search

Generative AI Policy

Background

Generative Artificial Intelligence (“GenAI”) is a type of artificial intelligence that uses algorithms trained on large datasets to create novel content—such as text, images, audio, or data—in response to user prompts. They are typically trained on massive datasets that include books, websites, news articles, Wikipedia, image-text pairs, and open-source code or audio—often drawn from publicly available or licensed sources. These diverse datasets help the models learn patterns and generate new content. Text-based generative AI systems learn patterns from enormous collections of written material and use those patterns to generate words and sentences that are likely to fit a user’s request.

Unlike predictive AI, which selects from preexisting responses, GenAI produces new outputs by identifying patterns in training data. While powerful, its results may reflect biases or inaccuracies present in the underlying data and should be critically evaluated.

Purpose

The purpose of this policy is to ensure the legal, ethical, and responsible use of GenAI tools at Wheaton College. These tools can create original content—such as text, images, audio, or data—based on large datasets. This policy outlines acceptable and prohibited uses, requirements for transparency, governance of approved tools, and safeguards to protect Wheaton’s mission, reputation, and compliance obligations.

Scope

This policy applies to all faculty, staff, students (only per below section, “Additional Guidelines on Student Use of Generative AI”), contractors, volunteers, and third parties with access to Wheaton College data or systems. It covers:

  • Internal self-developed GenAI models;
  • Third-party GenAI models deployed by Wheaton; and
  • Publicly available GenAI applications (e.g., ChatGPT, Gemini, Claude, Grok, DeepSeek) used for College purposes.

This policy applies whether GenAI is accessed on College-owned or personal devices.

Acceptable Use

Acceptable uses of GenAI at Wheaton College include using approved tools to support job responsibilities and enhance productivity, creating non-confidential administrative content such as orientation materials and signage, performing general text editing, brainstorming, and summarizing non-sensitive public information.

Prohibited Uses

GenAI tools must not be used to do the following:

  • Share confidential College information, including any personal or sensitive data about students, employees, or third parties—inputs and queries must not include the name of the College, the name of any person, or any personal or confidential information.
  • Analyze, debug, scan for errors, or otherwise review College documents or existing College software code containing confidential, sensitive, or protected information, as such actions could unintentionally disclose proprietary data to third-party platforms.
  • Engage in illegal activities, including hacking, copyright infringement, or unauthorized disclosure of confidential information.
  • Harass or discriminate another person based on race, gender, age, religion, or any other protected characteristic.
  • Generate inappropriate content such as hate speech, pornography, graphic violence, unwelcome altered images of any person, or any content that is inconsistent with the Wheaton College Statement of Faith or Community Covenant.
  • Impersonate any individual or entity or misrepresent Wheaton College.
  • Interfere with or disrupt the performance of Wheaton College’s systems, processes, or networks.
  • Engage in workplace behavior or conduct that violates any other Employee Handbook Policies.

The use of GenAI tools poses potential legal, commercial, and reputational risks involving contractual obligations, intellectual property rights, privacy, copyright infringement, and quality assurance. Specific contractual risks include:

  • Legal limitations on the sharing of confidential College information with third parties, including student records protected by privacy laws; and
  • Contractual agreements that restrict the purposes for which third-party data may be utilized. 

Licensing and Approvals

Use of GenAI platforms is governed by the platform's Terms of Use, Sharing & Publication Policy, Content Policy, and Usage Policies, which employees must review prior to use.

Employees may use free versions of Generative AI platforms without supervisor approval for tasks including:

  • Creation of administrative content such as non-confidential orientation materials, posters, or signage;
  • Answering routine questions and generating reports that summarize large volumes of nonsensitive, non-confidential, and non-legally regulated text or data; and
  • Editing text to improve grammar or to better convey a message with proper removal of confidential information.

Supervisor approval is required for use of free versions when:

  • Generating software code;
  • Producing research support content intended for external publication;
  • Accessing or producing content concerning global events occurring after 2021, given known limitations in datasets; and/or
  • There is any question about the sensitivity, confidentiality or legal implications of the text or data being utilized.

Paid versions of Generative AI require:

  • Submission and approval of a business case by the divisional head, copied to the Chief Information Officer (CIO); and
  • Departmental assumption of licensing costs.

Compliance and Accountability

Adherence to this policy is mandatory for all employees and students. For employees, violations may result in loss of access to Generative AI platforms or disciplinary action, up to and including termination. Alleged student violations of the Generative AI Policy will be addressed through the Student Conduct Policy in the Student Handbook.

Additional Guidance on Copyright and Intellectual Property

According to the U.S. Copyright Office, content generated by AI may not qualify for copyright protection, as it is not authored by a human being. Furthermore, there is a potential legal risk that Generative AI-generated content might be considered derivative work, potentially infringing on the copyright of training datasets if significant similarities exist.

Verification of Outputs and Transparency

All AI-generated content must be carefully reviewed to ensure accuracy—avoiding false outputs or “hallucinations”—as well as to identify and eliminate any bias or discrimination. When GenAI output forms a substantial basis of a final deliverable without significant human modification, disclosure must be made in the final material. Such disclosure is also required whenever applicable law mandates it.

Additional Guidelines on Student Use of Generative AI

No employee or student may be compelled to use Generative AI tools requiring the submission of personally identifiable information (PII) or other protected data to third-party providers. Students specifically are prohibited from entering PII and FERPA-protected information into Generative AI tools. For questions regarding intellectual property rights of content created with Generative AI, contact the College’s General Counsel at legal.and.risk@wheaton.edu or 630.752.5496.

Return to Handbook Table of Contents