The sale of fake Indigenous art has plagued Australia for decades, but a new accusation is adding a modern twist to an old injury. According to reporting in The Guardian, an Australian artist has warned that generative AI systems are hoovering up Indigenous Australian artwork and churning out cheap imitations, which the artist likened to the low-quality mass-produced goods sold on budget shopping platforms, describing the output as “Temu slop”.
The phrase is deliberately provocative, and it captures a genuine anxiety running through the country’s creative sector. For First Nations artists, the concern is not simply that machines can mimic a visual style. It is that the training data behind those machines was almost certainly harvested without permission, without payment and without any understanding of the cultural rules that govern who is allowed to paint what, and why.
An old problem meets a new technology
Indigenous art is one of Australia’s most distinctive cultural exports, and it has long been a target for exploitation. Successive inquiries have documented a trade in fake or inauthentic “Aboriginal-style” products, from boomerangs to didgeridoos to printed canvases, much of it manufactured overseas and sold to tourists who believe they are buying the real thing. A landmark federal parliamentary inquiry in 2018 found that up to three in four Indigenous-style souvenirs on sale carried no connection to any Aboriginal or Torres Strait Islander artist at all.
Generative AI threatens to supercharge that problem. Where a counterfeit canvas requires someone to physically print and ship it, an image model can produce an endless stream of convincing “dot painting” or “rarrk” style images in seconds, at almost no cost, and with no human hand involved. The images can then be slapped onto merchandise, used in marketing, or sold as digital files, all without a cent flowing back to the communities whose visual languages were scraped to build the tool in the first place.
The deeper injury is cultural rather than purely commercial. In many First Nations traditions, the right to depict particular stories, symbols and designs belongs to specific people and is passed down through kinship and Country. A machine that remixes those elements on demand does not know, and cannot respect, that a given design may be sacred, restricted, or simply not the maker’s to use. That is a form of harm that copyright law, which is built around individual authorship and commercial rights, was never designed to address.
Two sides of the argument
Artists and advocates argue that the current situation amounts to theft on an industrial scale. Their case is straightforward: the models could not produce Indigenous-style imagery unless they had been trained on real Indigenous art, that training happened without consent, and the output now competes directly with living artists while stripping their work of its meaning. Groups such as the Indigenous Art Code, which has spent years pushing for ethical standards in the sector, have warned that AI-generated imitations risk flooding the market and undercutting the authentic works that sustain remote communities.
The technology industry tends to frame the issue differently. Developers of large image models generally argue that training on publicly available material is permissible, that outputs are “new” creations rather than copies, and that style itself cannot be owned. Many point to the broader utility of the tools and to the practical difficulty of filtering vast training sets after the fact. That defence may hold up in parts of the world with permissive copyright regimes, but it sits awkwardly against Australian expectations around consent and cultural authority, and it does nothing to answer the question of who should be paid when a community’s visual heritage becomes raw material for a commercial product.
Why this matters for Australia
Australia is in the middle of a sustained debate about how to govern AI, and the treatment of Indigenous art lands squarely in the middle of it. The federal government has been weighing mandatory guardrails for high-risk AI uses, while the creative sector has pressed for stronger protections around how Australian work is used to train commercial systems. First Nations creators arguably have the strongest claim of all, because the harm they face is not only economic but cultural, and because existing copyright law offers them so little cover.
There is also a growing policy conversation about Indigenous Cultural and Intellectual Property, often shortened to ICIP, which seeks to recognise community ownership of traditional knowledge and cultural expression in ways that standard intellectual property law does not. Advocates have argued for years that ICIP principles should be written into law. The arrival of generative AI gives that argument fresh urgency, because the technology makes misappropriation faster, cheaper and harder to trace than ever before.
The economic stakes are significant. The Indigenous visual arts sector supports thousands of artists, many of them in remote communities where art centres are among the few sources of independent income. If AI-generated imitations erode the market or the premium that authenticity commands, the damage will fall hardest on the people least able to absorb it. That makes this a question of economic self-determination as much as one of copyright.
What happens next
Pressure is likely to build on several fronts at once. Expect renewed calls for AI developers to disclose what their models were trained on, and for a licensing or consent regime that would require permission and payment before Indigenous art can be used as training data. The push for stand-alone ICIP legislation, already on the agenda, will almost certainly gather momentum as cases like this one surface. There may also be moves to strengthen authenticity labelling, so that consumers can distinguish genuine First Nations work from machine-made imitations, building on earlier efforts to crack down on fake souvenirs.
None of that will arrive overnight, and the technology is moving faster than the law. For now, the artist’s blunt warning serves its purpose, reframing a slick new technology as the latest chapter in a very old story about who gets to profit from Indigenous culture, and who gets left out. The question for policymakers is whether Australia treats this as an unavoidable side-effect of innovation, or as a harm serious enough to regulate.
Sources: The Guardian.
