If an artist uses AI, should it affect their copyright?
Sunday, 8 March 2026
The Government says it is closely monitoring global developments in copyright law, as a New Zealand researcher argues art created using AI should receive a lower level of copyright protection than works made without the technology.
If an AI model is used to come up with a chapter for a novel based on an outline of the plot, that means the model is the thing putting flesh on the bone, so to speak, says Dr Joshua Yuvaraj, a senior lecturer at the University of Auckland whose research specialises in intellectual property law.
Copyright currently applies automatically when an original work like a book, painting or film is created. In most cases it lasts for 50 years after its creator dies ‒ soon to be extended to 70 years following the ratification of trade agreements with the European Union and the United Kingdom.
This means, with some exceptions, people cannot freely reproduce, publish or do other things with books, songs and other types of creative work without licensing.
Read more:
NZ authors among global payout out after AI chatbot trained on pirated books
‘Heartbroken and powerless’: Wairarapa author caught up in Meta’s book copyright scandal
Yuvaraj says if a machine does some or all of the creative labour, there is potentially leeway to pursue a shorter copyright term for those works.
However, if the law considers AI a tool like graphic design software, cameras or stationery, then in theory, works that are partially or wholly made using AI should be given the same scope and duration as copyright, Yuvaraj says.
A shorter term would mean content would free up much quicker for use: people would be able to reproduce and riff off those works without fear of copyright infringement.
Right now, the law says a work has to be original, which means in New Zealand there needs to be more than minimal skill or labour applied to it.
We have not yet had a test case about the use of AI prompts, but the US Copyright Office has refused to register AI-generated works, even when there was extensive prompting.
Yuvaraj says if there are no protections for AI-generated works, the country risks not rewarding considerable creative labour in the form of detailed prompting.
On the other hand, too much protection for AI-generated work could devalue human creativity.
“People might not exercise their brains when they can get the same protection for a fraction of the labour using AI. Getting the balance right is crucial.”
He suggests existing computer-generated work copyright provisions should be adapted, so they more directly apply to AI-generated work.
They allow for protection for works where there is no human author, but have been criticised and not widely used.
The time for protection Yuvaraj proposes for works that use AI is 50 years from creation, rather than life of the author plus 50 years, to reflect the lopsided balance in cognitive labour, and to acknowledge the role of the human.
Commerce and Consumer Affairs Minister Scott Simpson says there have been significant technology advances since the last major Copyright Act review in 2008.
While another review began in 2018, the process stalled and has not yet finished.
It is a timely opportunity to make sure the act is fit for purpose, Simpson says.
So far, very few countries have implemented specific changes to their copyright laws to address the development and use of AI tools. However, Australia and the United Kingdom are considering options.
Aotearoa has until mid-2028 to update some of its copyright laws, as part of the New Zealand-European Union free trade agreement.
The artist’s viewpoint
Auckland-based Māori singer and producer Huia Hamon says she chooses not to use AI in her creative music processes right now, other than for spelling correction and some grunt databasing work.
She feels a new provision of law, in which works that use AI could hypothetically receive less copyright protection, is fair.
Hamon hopes audiences will value music differently if they know AI was involved in creating it, but she is unsure if the public would be able to tell by ear.
“It is much cheaper to use AI to write music, but it’s no longer yours when you bring AI into the process. If you need to have AI to write music with you, you’re not a musician; you’re a DJ stealing other people’s tunes and saying it’s your sound.”
In her view, AI undermines the creative process and all the psychological benefits of practising art. “We derive ideas and inspiration from nature, conversations, DNA. These are things AI will never be able to experience.”
While AI has certain benefits, Hamon says because it is now “easy” to make music using the technology (that was illegally trained on others’ work), it is also promoting the Dunning-Kruger effect - a cognitive bias in which people with limited knowledge or competence greatly overestimate their own skill.
Conceptual artist James Ford says AI is just another tool in the creative process. However, due to its natural language interface and that you can chat to it and tweak prompts, sometimes “it feels like you are collaborating”.
Ford has integrated AI-generated imagery into some of his artworks, but says intentional prompting, the composition of the final artwork, use of non-AI generated text, and the fact the final artwork is a physical object with hand-painted elements, meant they would likely fall under existing copyright law, which stipulates human skill, labour and judgment.
“Questions of authorship and ownership have always been present in my work,” he says.
“These questions have been around for a long time in art, most obviously with the use of mass-production printing made famous by [Andy] Warhol, or further back in art history with overuse of assistants; or my favourite example, when John Baldessari commissioned painters and then openly displayed their work as his own.”