AI Use Standards & Safety Guide

Emerson College is embracing the future of Artificial Intelligence with curiosity, creativity, and care, always guided by the belief that creativity is a deeply human endeavor. While AI can be a powerful tool, it is not a substitute for imagination, judgment, or ethical responsibility. This guide supports thoughtful and transparent use of AI by staff with the goal of enhancing our work. We recognize that AI isn’t for everyone, and we encourage respectful dialogue and cautious exploration. Our goal is to support experimentation and innovation while upholding the values of creativity, inclusion, curiosity, equity, expression, and collaboration.

See also:

 

Using AI at Emerson: What’s Encouraged

AI can be an incredible assistant for:

  • Brainstorming: Generating ideas, outlines, or first drafts for creative or administrative work.

  • Summarizing: Condensing long documents, transcripts, or meeting notes into digestible insights.

  • Writing & Editing: Drafting memos, reports, presentations, or emails (as long as you review and personalize the results).

  • Data Analysis: Assisting with early insights into large datasets or reports (with validation).

  • Visual Design: Experimenting with image generation and layout concepts for non-sensitive creative work.

  • Learning: Exploring unfamiliar topics, coding, or concepts with guided AI assistance.

Remember: AI is a collaborator, not a decision-maker. To use it effectively, you need to stay in control. Think of AI as a partner you're working with, not a substitute doing the work for you.

Principles for Responsible Use

Disclose AI Output When Appropriate

If you use AI-generated content that contributes substantial, original language or ideas, particularly things you wouldn't otherwise be able to produce without it, and include that content in materials shared with others, you should clearly disclose it. Include:

  • The name of the AI tool

  • The date the content was generated

If you used AI to brainstorm, refine, or improve content you developed yourself, similar to how you might bounce ideas off a colleague, no disclosure is necessary.

Augment, Don’t Replace

Creativity, communication, and critical thinking are central to our work at Emerson. While AI can help you move faster or explore new directions, it shouldn’t take the wheel.

AI can help you do your job, but it shouldn’t do your job for you. AI should not:

  • Fully draft, approve, and send communications without your input

  • Attend or summarize meetings independently, without you having been present

  • Substitute for job duties. Use it to get started, generate a rough draft, or automate low-risk repetitive tasks—but always review the results with care.

Validate Data, Watch for Bias

AI is prone to hallucinations, bias, and outdated info. Question outputs, compare responses, and use external validation.

Protect Privacy and Data

Per the Data Governance Policy, do not enter:

  • Personally identifiable information (PII)

  • Confidential College data

  • Sensitive or restricted data Even small bits of info, when combined, can create privacy risks. Only share de-identified, non-confidential data.

Note: For Emerson-managed applications, such as Google Gemini and NotebookLM, your conversations are private and not visible to Emerson administrators.

Use Caution with Novel or Technical Tasks

If AI helps you write code, design workflows, or handle content outside your expertise, consult a colleague before using it. For example:

  • AI-written code should not be deployed unless reviewed by technical staff

  • AI-generated visuals must be appropriate, legal (i.e., don't use protected IP), and free of harmful bias

Creative Use

You can explore AI for creative projects like artwork, video concepts, or parody, while respecting the creative process. Of note, the College provides access to the entire Adobe Creative Cloud, which is rich with creative AI features.

Ensure that you label your work clearly if it uses deepfake, voice cloning, or similar technologies, and ensure lawful use of likeness.

Be thoughtful about how AI influences your artistic voice, and when in doubt, let your own ideas take precedence.

Use Reputable Tools

Stick to vetted tools or platforms approved by Emerson. Tools from lesser-known companies may not meet our privacy, security, or ethical standards.

See our list of vetted and authorized AI applications.

AI at the Platform Level

Many tools used at Emerson, like Zoom, Adobe, and Google, include AI features. Here’s what to know:

Procurement & Enablement

Before a department enables or purchases a new AI tool for broader use, it must undergo a security review per the Procurement Policy, just as with any application that will store College data. The Procurement Office will coordinate this process in collaboration with the Information Security team.

Security Standards

We will review AI tools for the same security standards we look for in all applications, including:

  • Multi-factor authentication

  • Encryption standards (NIST SSL/TLS)

  • Data minimization practices

  • Audit logging for security incident detection

Training AI with College Data

We will review new requests for AI tools to ensure that:

  • College data is only used to train AI if the model is fully controlled by Emerson and only improves our local instance.

  • Emerson data is not used to train a public or external model.

How to Get the Most Out of LLMs

  • Be Descriptive: Instead of "Write an email," try "Draft a professional email informing my team that the deadline for the project has moved to Friday."
  • Iterate and Drive the Conversation: AI is a collaborator and makes suggestions: you are the final decision-maker. Refine responses until they meet your needs. Verify information, cross-check facts, and ensure responses align with your expertise.
  • Request Citations for Verification. If an AI provides factual claims or data, and you are uncertain of its accuracy, ask it to provide links or references. Keep in mind that AI-generated citations may not always be valid, so verify sources independently when necessary.
  • Upload Context When Needed: AI can process documents to provide better responses. Upload project charters, agreements, or reports for more insightful analysis.
  • LLMs Are Not Search Engines. Unlike Google, AI models do not pull direct information from the web in real time. Instead, they generate responses based on patterns in their training data. Treat AI as a conversation partner rather than a search tool, and refine prompts accordingly.
  • The Best Way to Get Good Is to Practice! The most effective way to learn AI is by experimenting with different types of tasks. Start with small projects—such as summarizing reports, drafting an email, or generating ideas—and gradually expand to more complex work. The more you use it, the better you’ll understand how to phrase prompts effectively.

Example Tasks

Here are some powerful ways to integrate AI into your workflow:

  • Organizing Thoughts: Ask AI to structure your scattered ideas into an outline or strategy document.

  • Analyzing Agreements & Project Proposals: AI can help identify risks, inconsistencies, or opportunities.

  • Drafting Documentation: Create user guides, policies, or training materials.

  • Ideating Creative Concepts: Brainstorm marketing campaigns, video scripts, or app ideas.

  • Summarizing Long Documents: Condense reports, articles, or transcripts into key takeaways, or rework reports for different types of audiences.

  • Writing Code: Generate HTML, Python, JavaScript, or other programming languages for web development or automation.

  • Producing Graphics: Describe an image, and AI can generate visual assets, such as icons, illustrations, or concept designs.

Summary of Do’s and Don’ts

For the most effective, safe, and legally compliant use of AI, follow these principles:

DO:

  • Use AI for writing help, ideas, and analysis

  • Disclose AI use in distributed content when it's substantial or central to the output

  • Question and verify AI responses

  • Stick to de-identified, non-sensitive data

  • Explore AI creatively with appropriate labels

  • Use Emerson-reviewed platforms

  • Use AI to support, not replace, your unique creativity and professional judgment

DON’T:

  • Don’t rely on AI to generate work that reflects your values or voice without your careful input

  • Rely on AI to fully replace your job duties

  • Enter sensitive, private, or confidential info

  • Use AI-generated code or processes you can’t validate

  • Enable new AI features without IT review

  • Use deepfake tools without clear context and permission

For questions, concerns, or AI tool requests, contact helpdesk@emerson.edu.

Articles in this section