ACCEPTABLE USE OF AI

Related Policy – Acceptable Use of AI

About This appendix

University Appendix

Understanding Positive Uses of AI

AI tools can enhance productivity and decision-making across various workplace tasks:

Content Creation and Communication

• Generate first drafts of emails, announcements, or reports
• Improve word choice and clarity in professional communications
• Create meeting agendas and documentation
• Draft presentation outlines and talking points

Research and Analysis

• Summarize lengthy articles or documents
• Identify key themes in research materials
• Generate initial literature reviews or background research
• Explain complex concepts in accessible language

Planning and Organization

• Develop project timelines and task lists
• Create workflow documentation templates
• Generate training outlines or procedure frameworks
• Brainstorm approaches to new initiatives

Technical Assistance

• Debug non-sensitive code
• Explain technical concepts or formulas
• Generate script templates for repetitive tasks
• Provide syntax examples for programming questions

Important Reminder: AI tools serve as starting points and assistants, not replacements for professional judgment. Always verify outputs, customize content for your specific needs, and maintain human oversight of all work products.

Evaluating Data Before AI Input

To comply with data protection requirements, employees must follow these steps before inputting information:

• Review information against prohibited data categories
• Identify personally identifiable information (PII) or sensitive content
• For Restricted data, remove all identifying details and generalize context
• Consult with the supervisor or the Office of Innovation & Technology if uncertain

Verifying AI Content

To achieve accuracy and quality, employees must complete the following steps before using AI outputs:

• Fact-check all factual claims, statistics, and citations against authoritative sources
• Confirm the content aligns with university policies and standards
• Search for potential copyright issues
• Customize and edit content substantially to meet task needs

Writing Effective Prompts

The quality of AI output depends heavily on prompt quality. Well-crafted prompts lead to more useful, accurate, and relevant results.

Be Clear and Specific

• State exactly what you need and what the AI should produce
• Include relevant details like tone, format, length, and purpose
• Example: Instead of “Write about meetings,” say “Draft a 200-word email inviting department staff to a budget planning meeting on March 15, emphasizing the importance of attendance.”

Provide Context

• Share background information to help the AI understand your needs
• Identify your audience and what they need to know
• Example: “I need to explain our new travel reimbursement process to faculty who are unfamiliar with the university’s expense system.”

Use Step-by-Step Instructions

• Break complex tasks into smaller, sequential steps
• Guide the AI through your desired process
• Example: “First, list three benefits of the new policy. Second, provide a specific example for each benefit. Third, write a concluding paragraph summarizing the overall impact.”

Specify Format and Style

• Request specific formats like bullet points, tables, or paragraphs
• Define the desired tone (formal, conversational, technical)
• Example: “Create a table comparing the old and new procedures with three columns: Task, Old Process, New Process.”

Iterate and Refine

• Treat AI interaction as a conversation
• Ask follow-up questions to improve results
• Request revisions: “Make this more concise” or “Add more technical detail”
• Experiment with different phrasings to find what works best

Common Prompt Patterns

Role-Based Prompts: “Act as a project manager creating a timeline for…” Comparison Prompts: “Compare the advantages and disadvantages of…” Explanation Prompts: “Explain [concept] to someone unfamiliar with [topic]…” Revision Prompts: “Revise this text to be more [formal/concise/detailed]…”

Tool Selection Considerations

Microsoft Copilot

• Available at no cost to university employees through Microsoft 365
• Access by logging in with your NCU credentials
• Integrated with Word, Excel, PowerPoint, and Outlook
• Critical: Must log in with NCU account to ensure data protection – personal Microsoft accounts do not provide university data safeguards

Other AI Tools

• Before using any AI tool for university work, consult with the Office of Innovation & Technology
• Some tools may require specific approval or licensing
• Free public versions of tools like ChatGPT may not meet university data protection requirements
• University-approved tools receive regular security reviews and contractual protections

Key Selection Factors

• Data protection and privacy controls
• Integration with existing university systems
• Cost and licensing requirements
• Compliance with university policies
• Technical support availability

Best Practices Summary

• Start with specific, detailed prompts rather than vague questions
• Include context about audience and purpose
• Use AI tools as collaborative resources, not authoritative sources
• Recognize that AI systems generate plausible but potentially false information (hallucinations)
• Always verify factual claims before relying on AI output
• Maintain human judgment and professional expertise in all work products
• Document your AI use when transparency is important for understanding your methodology