Copyright in Gen-AI outputs
In Australia works can generally only be afforded copyright protection if there is a human author who contributed 'independent intellectual effort'. If you rely on Gen-AI to create your text/images, it is unlikely to be afforded copyright protection. The Government recently established the Copyright and Artificial Intelligence Reference Group (CAIRG) to consider whether and when AI-generated works should receive copyright protection. However, at present there is no clear threshold for when the 'independent intellectual effort' requirement is met in relation to using AI tools.
Copyright in Gen-AI inputs
To train Gen-AI models, developers scrape the web and copy large amounts of information. Often this occurs without the rights holders permission. In Australia there are currently no copyright exceptions for data mining or machine learning. If Gen-AI training has utilised material that infringes copyright and the work you create can be traced to this material, you could be issued with a takedown or infringement notice from the copyright owner.
Remember,
- uploading material from library databases such as journal article PDFs into Gen-AI tools will also breach licencing terms and conditions.
- uploading course content including assessment questions will breach the University's intellectual property policy.
Publisher policies & Gen-AI
If you are writing with a view to publish, check your publishers' policies before using any Gen-AI tools. All major academic publishers will not allow Gen-AI to be listed as an author, as they cannot be held accountable for ethical scholarship or inaccuracies. Transparency about the use of Gen-AI tools is a recurrent principle in publisher policies and you should not only document what Gen-AI tool you have used but also how you used them. Notre Dame's referencing guides include information on citing the use of Gen-AI tools.
See also
