A millennia ago, ancient Tamil dynasties ruled over vast empires straddling the Bay of Bengal. These trading civilisations formed a network state that was defined not by physical territory but by a distributed economic system, a shared culture, and the circulation of their currency.

Today, in what is now Sri Lanka, remnants of these ancient kingdoms are contested by a majoritarian government that instrumentalises archeology to assert its claims on the Tamil homeland of Eelam in the North and East of the island.

For three decades, Tamil Eelam was self-governed by a liberation movement that, in the early years of the World Wide Web, used the internet to coordinate a globally distributed parallel economic system amongst the Tamil diaspora, until 2009 when that de facto state was wiped out by the Sri Lankan government.

Interested in the challenges of narrating the losing side of a conflict when the winners invariably write history, Kulendran Thomas – whose family is from that nation that no longer exists – intervenes into a space of politicized archeological fiction. The artist’s newly commissioned work for ‘Nations by Artists’ is made using machine learning algorithms to excavate ancient coins from the subconscious of the internet.

Neural networks such as OpenAI’s natural language processing algorithms, that have ingested vast parts of all the world’s presently digitized information, are now able to generate new images based solely on linguistic prompts. Communicating with this type of network through ancient historical stories, Kulendran Thomas and his collaborators conjure from the internet a hallucination of three coins that have not previously existed. The artist’s multidisciplinary studio then sculpts these images into 3D digital models using an array of neural networks to synthesize detail that wasn’t ever there, resulting in hyperreal images that test the relationship between photography, archeology, and the historical narratives by which nations are built.