Generative AI (Gen-AI) is a type of artificial intelligence that has learned and then imitates patterns in large datasets to create new content such as text, images, and video based on user 'prompts'. Gen-AI tools can be used in variety of ways, including for study or research purposes - to synthesise, create, refine, test, inspire or modify.
Topic
Responsible use of Gen-AI tools
03 July 2024
Learn how to use Gen-AI tools responsibly.
HOW TO APPLY
Considering using Gen-AI in your work, research, or study?
Read this first:
Current Gen-AIs integrate technologies Large Language Models with web search capability to generate text. The underlying models have learned patterns in language and draw on these to construct their responses. While Gen-AI can summarize text, answer questions, and engage in human-like conversation, it does not "understand" the content or ideas in the way that humans do.
The quality of Gen-AI's responses is largely dependent on the precision of your prompts and the information it has access to. The more specific and well-defined your inputs, the better the results you can expect. However, it is also important to check the accuracy of the responses you receive
It is important to consider the following when using a Gen-AI:
Ethical Considerations: Use Gen-AI responsibly by considering the potential impact of its outputs and taking responsibility for them. Be mindful of biases related to gender, race, or social assumptions, and ensure that the generated content is accurate and relevant.
Data Integrity and Confidentiality: Exercise caution when inputting data into Gen-AI tools, particularly when you don't understand how the information will be used. Avoid sharing sensitive, confidential, or intellectual property-related information with Gen-AIs you don't trust, as it could be retained or misused. Stick to UC-endorsed tools like Microsoft Copilot with Data Protection, which adhere to privacy and security standards.
Intellectual Property and Copyright: Avoid disclosing others' intellectual property into Gen-AI tools whose privacy settings you haven't verified, as the information could become public. Before using, publishing, or sharing AI-generated content, consider whether doing so would violate copyright or intellectual property rights.
Accuracy and Reliability: Gen-AI may produce incorrect or misleading content. Always verify the outputs by cross-referencing with reliable sources, checking references, applying your own expertise or seeking second opinions. This helps ensure the information is accurate and trustworthy.
Assess data confidentiality carefully
The University has four data classifications:
public
in-confidence
sensitive
special
You should only use Gen-AI tools that are appropriate for the type of data you’re working with.
A very simple overview of how you would adhere to this might look like:
Special data
Information that has specific handling requirements, security controls or guidelines that must be followed to manage it.
We do not recommend using any Gen-AI tools.
Sensitive or in-confidence data
Information that needs to stay private for personal or organisational reasons. Using the University-endorsed tools means that your data will be erased after you finish your session and not be used to train the AI.
Microsoft Copilot with Data Protection when logged in with the University of Canterbury account.
Look for the 'Protected' symbol:
Public data
Information that is freely available. The University doesn't mind if AIs know this information or share it with others. It is freely available anyway.
Any Gen-AI tool.
My own personal data
Information about yourself. It is your choice who you share it with and you take the risk sharing it.
Any Gen-AI tool you feel comfortable sharing your information with.
UC staff
Find out more about information management and protection on our staff intranet (Staff only).
