Generative AI Guidance

What is generative AI?

Generative Artificial Intelligence (AI) is technology that helps machines perform tasks that usually require human intelligence. This includes understanding language, recognising speech, solving problems and generating content.

Large Language Models (LLMs), such as Microsoft Copilot and ChatGPT, can help with tasks like writing, summarising and research.

These tools can improve efficiency and support many everyday work activities. They are most effective when used thoughtfully and with human oversight.

The University recognises both the benefits and risks of generative AI. It is committed to using these tools responsibly, ethically and securely.

Who this guidance applies to

This guidance applies to all employees, contractors and volunteers who use generative AI for work-related tasks, such as drafting emails, summarising documents and preparing reports.

Guidance for students is covered separately under academic integrity and assessment policies.

Approved AI tools

The University provides access to Microsoft Copilot through Microsoft 365. This tool meets University standards for security, compliance and data protection.

Other tools, such as ChatGPT, Google Gemini, Claude and Perplexity AI, do not offer the same level of protection.

Use University-approved tools for official work wherever possible.

Benefits of generative AI

Generative AI can help improve efficiency, quality and accessibility across a range of tasks.

Efficiency and time savings – Speeds up drafting emails, reports and summaries.
Quality and consistency – Suggests clear structure and appropriate tone.
Audience adaptation – Rewrites content for different audiences and formats.
Accessibility support – Simplifies language and improves understanding.
Insights and decision support – Highlights key points and risks in long documents.
Knowledge discovery and search – Helps locate and extract information from approved sources.
Procedures and onboarding – Helps draft procedures, checklists and onboarding materials.
Creativity and brainstorming – Generates ideas, outlines and starting points.

Risks of generative AI

Understanding these risks helps you use AI safely and responsibly in your work.

Bias and inaccuracy – Outputs may contain errors or reflect biased language.
Hallucinations – AI may generate information that is incorrect, misleading or wrongly attributed.
Data privacy – Sensitive or personal information may be exposed if entered into AI tools.
Security risks – Unapproved AI tools may introduce security or data vulnerabilities.
Intellectual property – Outputs may include content that is copyrighted or restricted.
Compliance – Use of AI may breach University policies or legal requirements.
Reputational impact – Inaccurate or inappropriate content may damage trust.
Environmental impact – AI tools can require significant computing power and energy.

Using AI responsibly

Use these principles to make sure AI is used safely, responsibly and effectively in your work.

Use AI to create first drafts, rather than final outputs.
Always review and refine content before sharing it.
Do not enter confidential or sensitive information into AI tools.
Take responsibility for any decisions supported by AI.
Be open about your use of AI where appropriate.
Keep records of prompts and outputs where needed.
Apply your professional judgement at all times.

Practical examples

These examples show how AI can support common tasks while keeping you in control of the final output.

Drafting communications – Improve structure and tone, then review carefully before sending.
Summarising documents – Extract key points and check accuracy against the original.
Creating presentations – Generate a clear structure, then fact-check all content.
Writing procedures – Draft standard operating procedures (SOPs) and verify them with subject experts.
Improving accessibility – Simplify language while preserving the original meaning.
Brainstorming ideas – Use AI as a starting point, then refine and develop your ideas.

Using AI at work: dos and don'ts

Do

Use AI to support your work, not replace your judgment.
Draft emails, reports and routine communications more efficiently.
Summarise long documents into clear, useful key points.
Generate ideas, outlines and starting points for tasks.
Adapt tone and readability for different audiences.
Use AI as a research aid and check information against trusted sources.
Create draft procedures or guides for review and approval.

Don't

Enter personal, confidential or sensitive data into unapproved tools.
Rely on AI outputs without checking their accuracy and completeness.
Make important decisions based only on AI-generated content.
Upload passwords, login details or restricted documents.
Create or share misleading, biased or inappropriate content.
Bypass University policies, governance or approval processes.
Publish content without a named person taking responsibility.

More detailed guidance

Expand the sections below for more detailed advice on using AI safely and responsibly.

Use University-approved tools and devices to help maintain security and compliance.

Do not enter passwords, access keys or confidential information into AI tools.
Avoid sharing personal data or unpublished research.
Use licensed tools, such as Microsoft Copilot, for work-related tasks wherever possible.
Treat any information entered into non-approved AI tools as if it could be publicly visible.

Always review AI-generated content before using or sharing it.

Check facts, figures and references for accuracy.
Be aware of potential bias or misleading phrasing.
Clearly acknowledge AI use where appropriate.
Keep a record of prompts and outputs where decisions may need to be reviewed.

Make sure any AI-generated content respects copyright and intellectual property rights.

Check whether content comes from copyrighted sources.
Avoid copying or reusing content without confirming usage rights.
Cite sources where appropriate.
Use AI outputs as a starting point, not a final product.

Follow appropriate governance and approval processes, especially for sensitive or high-impact tasks.

Seek guidance before using non-approved AI tools.
Complete risk assessments where required.
Involve relevant teams for complex or high-risk use cases.
Follow approval processes before publishing or sharing outputs.

Use AI in a way that is responsible and aligned with University values.

Consider whether AI is necessary for the task.
Avoid excessive or unnecessary use of AI tools.
Be mindful of the environmental impact of high computational usage.
Choose simpler methods where they are sufficient.

Quick guide to AI risks


Risk	What could happen	What to do
Bias and inaccuracy	Incorrect or biased outputs	Review and verify facts
Hallucinations	Fabricated or misattributed information	Cross-check references
Data privacy	Exposure of personal or sensitive data	Avoid personal data; use licensed tools
Intellectual property	Copyright or licensing risks	Confirm usage rights
Compliance	Breaches of policy or regulations	Follow governance procedures
Reputational risk	Loss of trust or credibility	Ensure oversight
Security risk	Security or data vulnerabilities	Use approved tools
Environmental impact	High energy consumption	Limit non-essential use

Further information and references

This page was developed with the support of generative AI.