How AI imagery could be used to develop fake archaeology

Yahoo

04-02-2025

Generative AI is often seen as the epitome of our times, and sometimes even as futuristic. We can use it to invent new art or technology, analyse emerging data, or simulate people, places and things. But interestingly, it is also having an impact on how we view the past.
AI imagery has already been used to illustrate popular articles, such as covering scientific discoveries about Neanderthals. It was employed to animate the Mesolithic period (from about 9,000 to 4,300 years ago) in a museum. TikTok users have adopted it to make realistic short videos about archaeology and history. It's even been used in a TV documentary about Stonehenge.
Yet there are many issues with using AI imagery in archaeology – some of which are also found more broadly within generative AI use. These include its environmental impact and the violation of intellectual property (using training data created by humans).
But others are more specific to archaeology. As an academic who has worked extensively on 'resurrecting' the past through digital technology, generative AI has both fascinating potential and enormous risk for archaeological misrepresentation.
Even before the use of AI, it was widely accepted within archaeology that visualisations of the past are highly fraught and should be treated with extreme caution. For example, archaeologist Stephanie Moser examined 550 reconstructions published in academic and popular texts on human evolution. Her review found highly biased depictions, such as only males hunting, making art and tools and performing rituals, while women were in more passive roles.
A similar study by Diane Gifford-Gonzalez revealed that 'not one of 231 depictions of prehistoric males shows a man touching a child, woman, or an older person of either sex … no child is ever shown doing useful work.' These reconstructions do not reflect scientists' nuanced understanding of the past. We know humans organised themselves in an incredible array of variety, with a multitude of gender roles and self-expression.
A recent DNA-based study, for example, showed that women were actually at the centre of societies in the iron age.
The stakes of representation in archaeology are high. For example, the hotly-debated, dark-skinned reconstruction of 'Cheddar Man', originally found in south-west England, was based on ancient DNA analysis. It made headlines for disrupting the perception that all human ancestors in the north were light-skinned.
This and similar controversies reveal the iconic power of reconstructions, their political implications, and their ability to shape our understanding of the past.
While the Cheddar Man reconstruction demonstrates that research is iterative, such reconstructions are sticky. They have profound visual legacies and are not easily supplanted when new data becomes available.
This is exacerbated as they are incorporated into generative AI data sets. Beyond the use of outmoded data, generative AI visualisations of the past can be extremely poor.
Even when more plausible details are included, they can be seamlessly integrated with other highly inaccurate elements. For example, it is impossible for viewers to disentangle the data-led from the so-called hallucinations (mistakes) produced by AI.
Highlighting uncertainty is of central importance and concern among archaeologists. Archaeological illustrator Simon James noted that reconstruction artists have used strategically placed clouds of smoke to obscure unknown elements.
As a digital archaeologist, I have made virtual reconstructions of many different sites and subjects. I know there is often estimation and guesswork involved in making holistic representations.
Indeed, photo-realistic accuracy is not always the paramount consideration in visualisation – particularly when exploring different hypotheses or addressing young audiences. But knowing what is backed by archaeological data and what is more speculative is key for authentic visual communication.
This is particularly important at a time when pseudoarchaeology is increasingly prevalent in popular media, such as the Ancient Apocalypse show on Netflix. The celebrity host and author Graham Hancock asserts there was a lost ice age civilisation of Atlantis, with advanced technology. But this claim has been thoroughly repudiated by archaeologists.
Arguably, hoaxes will be much easier to perpetuate using generative AI. Beyond the high potential for misinformation about archaeology, the use of generative AI for archaeological visualisations can actually be harmful for archaeological knowledge production.
My research has shown that crafting reconstructions and illustrations in archaeology is incredibly important for understanding and interpreting the past. Creating visualisations based on science – and indeed soundscapes, smellscapes and other interpretations based on multiple senses – is very helpful for generating new questions.
Drawing allows archaeologists to create more detailed mental models and therefore a better understanding of archaeological remains. By delegating this creation to AI, archaeologists lose a powerful tool for knowledge generation. Moreover, my collaborative work with artists has demonstrated the intriguing possibilities that creative approaches open up to tell new stories about the past.
Even with all of these problems, I encourage an engaged, critical, applied approach to understanding the impact of digital technologies on our investigation of the past. And this includes exploring the uses of generative AI for archaeological visualisation.
Archaeologists and non-specialists are able to leverage generative AI to creatively produce interpretive media. Indeed, some archaeologists are already exploring AI to generate hypotheses about ancient life. And we are teaching critical uses of AI to our archaeology students.
But what remains imperative is that archaeologists engage with and critique all visualisations – both those created by generative AI and using other media.
This article is republished from The Conversation under a Creative Commons license. Read the original article.
Colleen Morgan does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.