Copilot Chat (Generative AI)
Copilot Chat is included with our current Microsoft 365 licence and is available to all UQ staff and students 18+. Copilot Chat for the 13–18-year-old cohort is rolling out in June 2025. It uses information from the web and any uploaded documents to assist with various tasks, such as:
searching for relevant information on the web
generating content such as summaries, presentations and code
rewriting, improving, or optimising your existing content
creating or drawing images based on your descriptions.
It is different to Microsoft 365 Copilot, which includes Copilot Chat and extends AI capabilities into Outlook, Word, Excel, PowerPoint, and Teams.
Access Copilot Chat
Copilot Chat is accessible via Microsoft Teams, the Microsoft 365 Copilot app and a web browser. The instructions that follow relates to accessing it via Microsoft Teams.
To access Copilot Chat via Microsoft Teams, follow these steps:
Open Microsoft Teams: Launch the Microsoft Teams app on your desktop or mobile device, or access Teams via a web browser.
Sign in with your UQ credentials, if you are not already signed in.
Find the Copilot app: if it is not available on the left-hand rail, select Apps and search for Copilot.
Select Add or Open to display Copilot Chat.
Pin Copilot app
If the Copilot app is not currently pinned on the left-hand rail, right-click Copilot and select Pin.
Privacy and protections
- Data is secured: data is encrypted at rest and in transit
- Your data is private
- Prompts and responses are logged, retained and available for audit, eDiscovery, and advanced Microsoft Purview capabilities. You can access previous Copilot chat for reference or to continue the chat
- Prompts and responses are not used to train foundation models
Academic Title Holders
Academic Title Holders do not have access to Copilot. This is due to Microsoft rolling out access to specific Microsoft licences only. This may change in the future.
Risks of using generative AI?
There are some risks and limitations that you should be aware of when using generative AI. As with any technology, staff should not use it in a way that breaches UQ policies, procedures, or codes of conduct. There are common risks and here are some guidelines to keep your data safe and enhance your experience.
- Accuracy of output: generative AI learns from large data sets of unknown (and/or unverified) sources, which can result in inaccurate or biased output. It may also provide incorrect or misleading information, or fail to understand the user’s intent or context.
For example, AI models may be prone to "hallucinations", where the AI tool might create outputs which are nonsensical or completely inaccurate. AI hallucination types can include content that appears trustworthy or authoritative - such as references to scholarly articles or detailed case studies - that are fabrications.
To mitigate the risks and enhance the experience of using generative AI, you should follow some best practices, such as:
- Human in the loop (HITL): You should always validate the accuracy of the output you receive from generative AI tools using independent sources; you may also wish to ask for sources for the content it provides and should independently review these sources to ensure that any inference or conclusion drawn from them is justified. You should also monitor and review its performance and provide feedback to improve its quality.
- Check the terms of service: You should check the privacy policy and terms of service and understand how it handles your data. Staff should not enter information classified as SENSITIVE or PROTECTED into generative AI tools.
- Authorisation: You should ensure that you have the necessary rights and consent to use the data you provide to genertive AI tools. For example, if you are entering personal information, you require the consent of the person whose information is being used.
- Product is generally available: Review the features, functionality, and reliability of generative AI tools.
Policies, procedures and licensing agreements
As with any tool available at UQ, all UQ staff and students using genertive AI must be aware of the relevant policies, procedures and licensing agreements that govern the content they are using or creating.
Frequently asked questions
Can I use AI to create images?
Yes, with the following considerations:
- You will have to credit the use of AI if you use the image anywhere e.g. An AI-generated image of an owl.
- These images and graphics will have no copyright protection, which means anyone else can use this content.
Can I use AI to create UQ sub-brands, logos and identities?
No. While you can use the functionality of prompting AI to mock-up hypothetical logos, none of the content can be protected under copyright.
UQ sub-brands, logos and identities must not be created without prior approval from Marketing & Communication via brand@uq.edu.au. For direction, refer to the brand decision table in the UQ Brand Guidelines [PDF].
The creation of names, sub-brands and trademarks are formalised through prior approval and consultation with Marketing and Communication with final approval from the Chief Marketing and Communication Officer and Vice-Chancellor.
Can I use AI to summarise UQ-provided textbooks or journal articles?
Refer to Policies and guidelines.
Can I use AI to generate content for UQ websites?
This is not advised. Outputs from generative AI are not eligible for copyright protection. As any content created by AI must be identified as such, this content can then be used freely by anyone else who wishes.
More information
If you would like to know more about the service, visit the Microsoft Copilot webpage.
If you have feedback or a question about Copilot or AI Chatbots, please share it with ITS.
Related content
Artificial Intelligence at UQ (staff-related guidance)
Artificial Intelligence Resource Page (Tertiary Education Quality and Standards Agency)