Artificial intelligence tools may be used to support drafting, editing, summarization, and accessibility improvements. AI must never be used to process confidential, protected, or sensitive information. All AI鈥慳ssisted content remains subject to human review, institutional policies, and A&F branding and ethical standards.
Policy: The use of artificial intelligence (AI) within A&F must comply with all UMB and CITS policies. Where appropriate, AI can be leveraged in developing all types of communications both within and external to A&F. Confidentiality must be prioritized when handling sensitive information.
Artificial Intelligence (AI) is a rapidly evolving technology that can enhance efficiency, creativity, and innovation in the workplace. While the capabilities of AI are expanding exponentially, the rules governing its usage should be familiar to all. All AI use and output are subject to the same core values, branding guidelines, and ethical considerations as all other communications.
Given the quickly evolving nature of AI, the information and guidance provided here is expected to change over time. These resources and tools are for guidance purposes only, and A&F employees are STRONGLY encouraged to seek out UMB-sponsored training programs for further education and use of AI tools at UMB.
Procedure:
For guidance on using AI, please refer to these UMB approved pages for answers:
- The UMB AI Governance policy can be found here.
- The official CITS overview of AI can be found here.
- UMB AI Guidelines can be found here.
- UMB approved AI tools can be found here.
- Information about using AI to assist with your writing communications can be found here.
Common Uses for AI
Several common uses for AI exist and which can be leveraged within the A&F work environment. AI is an effective tool in:
- Brainstorming ideas or outlines
- Drafting early versions of routine communications, including emails
- Summarizing long documents, emails, etc., for internal use
- Improving clarity, tone, or accessibility of text
- Generative alternative phrasing or headlines
Here are some additional ideas to keep in mind when using AI.
The prompt will shape the output.
How you speak to AI will determine how it will respond. Specific questions will often generate specific answers. If you’re brainstorming or looking for creative inspiration, open ended, less concrete asks can generate better results. Simply, the quality of what you enter into AI as prompt will influence the quality of the AI tool output.
AI is not a vending machine for facts.
AI models can and do generates incorrect, nonsensical, or fabricated information. This is known as an AI hallucination and data scientists predict it will be impossible to eliminate. This means every AI output should be confirmed or verified. AI will force itself to answer questions and present information as factual, even when it’s just guessing.
AI really wants to please you.
It can be useful to think of AI like a brilliant intern. It will do what you ask of it, it will work amazing fast, and it’s really good with words. The downside is that it’s going to try and do things for you whether it’s able to or not. While it has just about all of the world’s information at its disposal, it has almost no wisdom of judgement. It’s always up to the user to apply human judgement.
Data, Privacy, and Risk
While AI is an optional, effective, and efficient tool, there are some critical guidelines to follow when using AI tools at UMB and within A&F. This list is not fully inclusive and AI users are encouraged to review the UMB and CITS guidelines and policies before consistently utilizing AI.
AI should never be used to handle confidential, protected, or sensitive data. Types of private information that should never be entered into AI tools, include, but are not limited to: personal identifying information (e.g., date of birth, social security number); student records, personnel information, protected research data or health information, etc.
- AI tools used behind the UMB firewall – e.g., CoPilot – have met institutional privacy and security standards.
- Public tools outside the UMB firewall may retain or reuse input data unless otherwise specified.
- When using AI at UMB, personnel are strongly encouraged either to utilize AI behind the UMB firewall or refrain from sharing sensitive information when using AI external to the UMB firewall.