Guidelines for Use of Artificial Intelligence (AI) Tools

The term AI refers to tools and systems that mimic human intelligence, such as understanding language, solving problems, and creating content (i.e.; generative AI). Examples of popular AI tools include OpenAI’s ChatGPT, Google Gemini, Grammarly, and Microsoft Copilot.

AI, most notably generative AI, is a rapidly evolving technology which is being incorporated into many online products and services. Generative AI goes beyond traditional AI systems by creating new data that resembles human-created content. While the application of AI technology is being rapidly adopted, the governance and security aspects of AI tools may not be maturely developed. Meaning, the information you provide an AI tool may be used in ways you did not intend or anticipate.

When using AI tools, it is important to be mindful of information security and data privacy, compliance, copyright, and academic integrity.

The following guidelines are based on existing TSUS and college policy.

Acquisition and Official Use of AI Tools

Faculty and staff seeking to use or acquire AI tools must adhere to college policies and standards regarding the acquisition of new IT products and services. Established college procurement processes ensure that IT products and services have sufficient compliance, privacy, and security protections to safeguard our faculty, staff, and students. While a number of authorized applicationsand services already in use at the college utilize AI, only one actual AI tool is approved for use.

Microsoft Office 365 Copilot

URL: https://m365.cloud.microsoft/chat/

M365 Copilot app is bundled with the college’s Microsoft Office 365 subscription. Copilot is built on OpenAI’s ChatGPT-4 architecture but has been customized and optimized by Microsoft. The college is currently licensed to use it as a standalone app that does not integrate with Microsoft Office. Information provided to Copilot is not shared externally and is not used for training.

• Data, Privacy, and Security for Microsoft 365 Copilot
• Enterprise Data Protection in Microsoft 365 Copilot and Copilot Chat
• Microsoft 365 Copilot Chat Privacy and Protections

Guidelines for use of AI tools

Safeguard Sensitive and Confidential Data: Do not enter sensitive or confidential information into AI tools not authorized by the college for such usage. Information shared with third-party AI tools could be exposed to unauthorized parties.

Use Responsibly: You are responsible for any content you produce or publish, including AI-generated content. AI-generated content can be inaccurate, misleading, or entirely fabricated (sometimes called “hallucinations”), or may unintentionally contain copyrighted, sensitive, or even confidential material. Review your AI-generated content before publication or sharing.

Maintain Academic Integrity: Students’ use of AI tools must adhere to college requirements regarding academic integrity found in college policies, handbooks, codes of conduct, and other such documents. Faculty should be clear with students in their courses regarding permitted uses, if any, of AI tools. Students are also encouraged to ask their instructors for clarification about permitted uses of AI tools as needed.

Cautions

• Some AI may be prohibited technology (e.g., Deepseek and Midjourney) and as such not allowed for use on college networks or college devices.
• Double-check factual claims and citations. AI can sometimes generate plausible but incorrect information.
  • Why is AI so bad at math
  • What are AI Hallucinations?
• Save important conversations or outputs elsewhere. AI chat histories aren't permanent.
• Remember that AI has limitations. It can't verify real-time information and may not be able to learn from past chats.
• Remember that AI may be biased. AI models may reflect biases in their training data, so use critical thinking when analyzing their results.
• Most AI will use data you provide to it to train its program. Be mindful of what information you provide and ensure you are following college policies on confidential and sensitive information.

Prompt Generation

• Be clear and specific about what you want.
• Provide step by step instructions. Break complex tasks into smaller steps to get more accurate responses.
• Provide context, such as relevant background information, tone of the output, and other details that will help make the output more accurate.
• Identify the type of output you want, such as an email, a formal letter, a blog post, or a checklist.
• Identify the format of the output, such as plain text, a Word document, or a PDF document.