Everything is made from carbon: soil, your body, a computer. We’ve just chosen to forget the family resemblance.
Maps did something specific: they drew a line around land and said, this counts, this doesn’t. The same instrument used to survey Dartmoor became the legal tool used to extract from it. Algorithms are doing the same job now, just faster, deciding whose knowledge counts as data and whose doesn’t.
That’s not a metaphor. Every large AI model needs a physical data warehouse to run. Some UK banks have already stopped lending for new ones, the power and cooling costs make them too expensive to build. The faster the AI, the more land and energy it takes to run. Speed has a physical cost, paid somewhere, by someone.
A wider, ongoing body of work lives at adahparris.art. These are the pieces that connect most directly to the work here.
Selected pieces
Working title: 'The Middle East'
Middle to whom?
A single strait carries a third of the world's fertilizer, a quarter of its oil, and the cables training its AI. Soil, DNA, and algorithm, converging on one stretch of water.
In focus
A single strait carries a third of the world's fertilizer, a quarter of its oil, and the cables training its AI. Soil, DNA, and algorithm, converging on one stretch of water. We're not in a technology crisis. We're in a listening crisis. What becomes possible when we remember that intelligences are entangled and animate?
No title yet. No image yet. Research in progress.
Before the Map: Ancestral Cosmologies, 1897 to Now
Ink, acrylic paint, and pencil on antique celestial atlas · 2026
This 1897 star chart decided which constellations deserved names. Nine points across it mark where an ancestral sky-tradition and a modern AI language project name the same place, sometimes in agreement, sometimes in open contradiction. Oshumare, the Yoruba rainbow serpent, stretches across both hemispheres, arriving in dream rather than research. Part of the Cartography Is Not Home series.
In focus
Nine points — Nine points across an ancestral sky. When a modern language model and an indigenous cosmos name the exact same geography differently, which system of naming is allowed to survive?
Oshumare — Oshumare arrived through a dream, not a survey. Every AI dataset has a blind spot shaped by its own methodology, a reality that lies beyond what an algorithm can scrape. What’s missing from yours?
Available to acquire, or as the starting point for a commissioned piece in a similar vein — see /acquire.
Before the Map: Ancestral Cosmologies, 1897 to Now
L3b1a + D4b1a2: Original Instructions
Mixed media · gold thread on original atlas page · 2025–2026
A 1897 North Polar Regions atlas page. Gold thread connects every leyline through the centre, through 90°N, through L3b1a + D4b1a2, the two haplogroups separated by 60,000 years of divergence that share a single East African maternal ancestor. Haiti, Guyana, Congo, Sudan, Palestine, Iran, Bangladesh, Western Sahara. Arrhenius's 1896 climate warning in the artist's hand at the top. CO₂ then: 295 parts per million. CO₂ now: 420. The thread does not radiate outward from a source. It passes through, accumulating at the centre with each pass. As DNA passes through generations. As mycelium passes through soil.
In focus
Palestine — Levantine and Ashkenazi lineages share common ancestral haplogroups, yet territorial claims seldom make room for shared roots. Today, that same lineage and demographic data feeds the algorithms automated systems use to categorise, target, and displace people online. The technology changes, but the core function remains: does the sorting get any gentler, or just faster?
Iran — Underground water channels, built three thousand years ago, still carry water the way DNA carries lineage. Passed through, not sourced from one point. The sanctions are recent. Which one was built to last?
Available to acquire, or as the starting point for a commissioned piece in a similar vein — see /acquire.
L3b1a + D4b1a2: Original Instructions, 1897 North Polar Regions atlas page with gold thread
Begin
Interested in commissioning new work, a gallery premiere, or an institutional partnership? Start here →