For decades, the idea that ancient astronauts visited Earth and left their mark on humanity was the stuff of late-night radio shows, low-budget cable television specials, and dog-eared paperbacks found at the back of used bookstores. It was a concept that, for most rational people, occupied a comfortable place on the shelf of harmless, guilty pleasures—entertaining science fiction that you didn’t have to take seriously. Within this peculiar niche, a recurring theme involved ancient rock art, with the flamboyant claims of pseudoarchaeologists often pointing to the extremely old paintings found across the Australian continent. Specifically, they pointed to the Wandjina figures—the striking, large-eyed, headless-looking ancestors of the Mowanjum people of the Kimberley region—and tried to pass them off as literal portraits of extraterrestrial visitors in spacesuits. For a long time, these were simply the ramblings of a fringe sector, easily dismissed by academics, historians, and most importantly, by the Indigenous custodians of these sacred sites, who viewed the claims as laughable, if not slightly aggravating, intrusions into their lived culture.
But the internet changed the digital tide, and now, a more formidable and dangerous force is threatening to turn these fringe musings into a confused mainstream reality. That force is Artificial Intelligence. Unlike the static nature of a paperback book or a television documentary, AI is dynamic, interactive, and uniquely capable of synthesizing information in a way that mimics human expertise—but without any underlying comprehension of truth. It is a technological machine that, when asked, “What does Aboriginal rock art depict?” will scrape the vast, messy archive of the internet and present a collection of responses. Because the modern web is a cacophony of voices where an academic paper published by a university carries the same digital weight as a conspiracy blog run from a basement, AI inevitably begins to treat these fringe theories as equal, or even more compelling, alternatives to the established anthropological consensus. The result is a digital regression, a return to a kind of algorithmic echo chamber where the absurd is presented with the same neutral confidence as the factual. This isn’t just a threat to intellectual accuracy; it is an existential cultural threat to the very communities whose history is being manipulated and digitally colonized.
To understand the gravity of this issue, we must first strip away the academic jargon and get to the heart of what these rock art sites actually are. For the Aboriginal peoples of Australia, the oldest continuous living culture on Earth, these sites are not “ancient artifacts” or “prehistoric relics.” They are the living, breathing, physical manifestations of the Dreaming—the time outside of time when ancestral spirits created the world. The Wandjina figures, for instance, are not drawings of aliens; they are the resting spirits of the great creators themselves. They are the entities responsible for bringing the rains, the tides, and the winds. The paintings are the physical guarantee that the spiritual energy of the ancestors remains in the land, looking out for the living community. When an Elder looks at these giant eyes and elongated faces, they are seeing the face of their grandfather’s grandfather, a testament to their own unbroken lineage tying them to the land for tens of thousands of years. To reduce these sacred beings to the level of extraterrestrial visitors is not just disinformation; it is a violent destruction of a living cultural identity. It removes the texture of spirituality and deep law and replaces it with a simplistic, colonial-era fantasy that says these people couldn’t have created such profound art on their own—it must have been spacemen. It is a modern continuation of the very colonial mindset that originally dismissed Indigenous people as primitive, needing external intervention to explain their complex achievements.
Now, consider how AI ingests and regurgitates this context. Large Language Models (LLMs) are, in essence, massive pattern-matching engines. They are trained on billions of parameters of human text, absorbing everything from peer-reviewed journals to subreddit threads. When a user types a query like “Explain the figures in the Wandjina cave art,” the AI is not consulting a database of verified facts. Instead, it is predicting, word by word, the most plausible sequence of text based on the statistical probabilities of the linguistic patterns it has absorbed. It attempts to mimic the style of an authoritative answer, often without the human ability to judge whether that answer contains deliberate misinformation or idle speculation. If a high-school student, a curious journalist, or a public-school teacher asks an AI chatbot about the origins of the rock art, and the AI returns a plausible, confident answer that suggests, “Some researchers theorize a connection to ancient astronauts,” the damage is already done. The AI has effectively laundered a fringe conspiracy theory, giving it a veneer of objectivity and neutrality it never possessed in the hands of a fringe blogger. The result is that the most vulnerable audience—people who are digitally literate but historically uninformed—will walk away with an incredibly distorted version of the truth, unaware that they weren’t reading an impartial encyclopedia but rather the aggregated conclusions of the internet’s most deranged corners.
This “mainstreaming” effect is the great danger of the AI age. Historically, a conspiracy had to be actively sought out. You had to buy the book, watch the cable special, or visit the specific website. This “filter bubble” kept the most harmful nonsense contained. AI destroys that containment entirely. It acts as an algorithmic bridge, connecting a mainstream query to a fringe answer without any warning lights flashing. It turns a participant into a passive consumer of misinformation. Furthermore, this is a two-way street. As AI generates this sloppy content, it gets written into blogs, websites, and social media posts. Then, newer AI models are trained on that same freshly generated data, creating a self-fulfilling feedback loop of falsehoods. The more the AI says it, the more the AI believes it. This matter is compounded by tools like AI-driven image generators, which can now create stunningly realistic images of “Aliens at Uluru” or “Spaceships passing over the Outback.” These images, circulating on social media, become decontextualized “proof” for the uninitiated. We risk a future where the line between the sacred, deeply documented human history of an Indigenous people is permanently blurred by the confabulated output of a machine that was never taught the difference between a sin and a stroke of code.
The true cost of this for Aboriginal communities is not merely rhetorical; it has tangible, devastating consequences. It directly undermines the authority of Indigenous peoples over their own heritage. Elders have spent decades fighting for sovereignty over their stories, their ancient knowledge, and their sacred sites. They have successfully argued that they are the sole custodians of the narrative of their own creation. When an AI presents the “ancient alien” theory as an alternative, it is literally erasing the human agency and the spiritual authority of a living people, replacing it with a technocratic, western fantasy. This isn’t merely an offense; it is a direct assault on their struggle for self-determination and reconciliation. Moreover, this misinformation emboldens a specific type of traveler: the “pseudo-archaeologist” or the treasure hunter. Historically, these theories have fueled the desecration of sacred sites by tourists and hobbyists who believe they are uncovering a “hidden truth” that governments and universities are hiding. If AI, accessed by millions of users daily, presents these theories as legitimate, it will inspire a new generation of visitors to seek out these remote rock art sites not to pay respect to the culture, but to hunt for evidence of “ancient astronauts.” They might touch, scratch, or photograph the art in a way that permanently damages these priceless, 7,000-year-old sites—all in the name of finding a fictional alien glyph. This is where the abstract threat of “misinformation” meets the concrete reality of environmental and cultural vandalism.
In facing this reality, we cannot simply throw our hands up and blame “the machine.” We must rely on a deeply humanized response. The remedy is a global commitment to digital literacy and, perhaps more importantly, a fierce recommitment to centering Indigenous voices as the primary sources of authority on Indigenous history. We must teach the next generation that not all information produced by a glowing screen is trustworthy, and that the quest for knowledge requires humility, particularly when the subject is not ours to interpret. We must push for AI models that are not just trained on volume, but are curated with an understanding of context, sovereignty, and cultural safety. This means holding tech companies accountable for the datasets they use, demanding that they incorporate stringent checks that flag content from Indigenous knowledge holders as higher priority than fringe conjecture. But fundamentally, the solution is to reconnect with the human chain of knowledge. The paintings are not data to be mined; it is a story to be spoken. The Wandjina figures are not mystery images to be decoded; they are grandparents to be listened to. We must choose to listen to the Elders, to the anthropologists who respect them, and to the documented history of the continent—not to the confident, soulless voice of an algorithm. If we lose this battle, we lose not just the truth about a rock painting, but the human right to own one’s own story. In the end, the fight against the AI mainstreaming of this nonsense is a fight for the soul of human history itself. It is a reminder that in a world increasingly built by machines, our humanity must remain at the helm, guiding the technology with empathy, respect, and a profound recognition of the living voices that came before us.