Human Resources
Generative Artificial Intelligence (AI) is a rapidly evolving technology with increasing popularity of generative AI chatbots (e.g., ChatGPT, Google Bard, etc.). While we want to encourage use of new and emerging technologies to work more productively, we also want users to understand the risks and limitations of generative AI chatbots to ensure they are used responsibly. It has become necessary, therefore, to put in place this policy for responsible use of these tools. This policy is likely to change, perhaps multiple times over the coming months. We suggest you bookmark this webpage and refer to it often.
This policy is intended to only govern administrative work and applies to all Union College employees (staff and faculty), including student employees. This policy does not govern the use of chatbots in teaching and scholarship.
Policy
Limited use of generative AI chatbots by employees (staff and faculty) and student employees is allowed while performing administrative work for Union College. While generative AI chatbots can be used to perform a variety of functions, this policy addresses only the use of a web-based interface to ask or “prompt” the chatbot in a conversational manner to find answers to questions or to generate written content.
Employees may use their Union College email address, credentials, or phone number to set up an account with these AI tools provided you understand and abide by the following:
Employees are responsible for any content they produce using generative AI chatbots.
There are high-stakes risks in using these tools, including uncertainty about who owns AI-generated content, and security and privacy concerns associated with inputting any confidential, sensitive, or private information. This includes information about prospective, current, and former students; job applicants and current and former employees; trustees, alumni, and donors; and other persons or parties affiliated with the College.
Content generated by AI tools cannot be relied upon, as the information may be inaccurate, outdated, misleading or—in some cases—entirely fabricated (also referred to as “hallucinations”), or may contain copyrighted material.
Following is a list of typical acceptable uses of generative AI chatbots provided that such use complies with this and all other applicable college policies:
To produce draft text for:
- Emails, memos, letters, or other forms of correspondence;
- Marketing copy, articles, social media posts, or announcements;
- Talking points and remarks for presentations, meetings, or public events;
- Job descriptions, job ads and postings, or onboarding plans; or
- Policies, procedures, or work plans;
To brainstorm ideas related to projects or assignments.
To summarize, outline, or sort internal and external information that is not confidential, sensitive, private, or personally identifiable.
To create spreadsheet formulas.
To translate text into other languages.
When a staff member or student employee wishes to use a chatbot for a novel purpose that is not listed above, their supervisor may ask them to put their request in writing and may consult with or refer their request to the appropriate college officials before a decision is made to approve or deny their request. Until such a request has been reviewed and approved, they are not to use a chatbot for the purpose they are requesting.
When using generative AI chatbots, do not copy and paste, type, or submit confidential Union College-related information or data into a generative AI chatbot. Employees must strip any such information or data before entering prompts into a chatbot.
When entering prompts into a chatbot, use placeholders instead of Union College-related information or data. For example, replace a reference to Union College with “college” or “organization.” For additional guidelines pertaining to proper handling of Union College data, refer to the College’s data classification policy, which can be found on the Information Technology Services website here.
Unedited text generated by a chatbot that will appear in a final work product and could be reasonably presumed to be original work must be properly cited as chatbot-generated. Chatbot-generated content must also be reviewed for accuracy before relying on it for work purposes and properly cited when used as a resource for factual information. If a reliable source cannot be found to verify factual information generated by the chatbot, that information cannot be used for work purposes.
As AI chatbots may produce content that is plagiarized from its knowledge base, including copyrighted works, no text generated from a chatbot will be eligible to have a Union College copyright or trademark. Refer to the College’s intellectual property policy in the Faculty Manual.
Violations of this policy may result in disciplinary action, up to and including termination of employment.
If you have questions or feedback about this policy, please contact HR.
Training Resources
Writing effective prompts is the key to using generative AI chatbots productively. This LinkedIn Learning course – How to research and write using generative AI tools (1 hour, 10 minutes) – provides an excellent overview.