■ SAFE RISK ■ Agriculture & Environment
Not replaced, though the collar on the lead animal already does work your grandfather did on skis. The herding survives; the surveillance part of it is largely gone.
“GPS tracks the herd. But navigating Sami traditions through Arctic tundra needs a human.”
Our AI replacement risk score — how we score jobs
Herding runs on a seasonal calendar rather than a working day: spring calving on the mountain, summer grazing dispersal, the autumn gathering and separation at the corral, winter feeding when ice locks the lichen under a crust. Between those, the work is snowmobiles, quad bikes, fence repair, predator loss reporting, negotiating grazing boundaries with neighbouring siida, and hours of watching animals to know which ones are yours and which are drifting.
GPS collars changed the economics of the job more than anything since the snowmobile. A herder can now see on a phone that four hundred animals have moved into a forestry block overnight instead of driving lines for two days to find out. Satellite and drone snow-cover data helps predict icing events. Slaughter logistics, subsidy claims, earmark registers and predator-compensation paperwork have all moved onto software. The scouting labour — the pure information-gathering — has genuinely collapsed.
What did not move is everything involving animals, land and law simultaneously. Reindeer are semi-wild; they do not respond to machinery the way herded sheep do, and moving a nervous herd across a road or a river takes people reading animal behaviour in real time. Calving losses, disease calls and lassoing at the corral are hands-on. Above that sits the political layer: grazing rights contested with mining companies, wind farms and forestry, an inheritance and kinship structure that decides who herds at all, and cultural knowledge — snow terminology, migration routes, marking traditions — that exists in families rather than datasets. The threats to this occupation are land use and climate, not algorithms.
Automatability: our editorial assessment of current and near-term AI capability
The technological shift already happened — collars and satellite snow data are standard, not futuristic. Nothing on the horizon to 2040 removes the human from calving, corral work or grazing-rights negotiation. The real disruption arriving this decade is ecological: unstable winters, ice crust events and land taken for wind and mining, which threaten the occupation far more than automation ever will.
Mostly it has helped. Collars removed days of searching per week and let herders react to a herd drifting toward a road or a forestry plantation before losses occur. The complaint you hear is cost and dependence on network coverage, not job displacement. The labour saved goes into corral work, paperwork and fence maintenance that never got done before.
Viable but squeezed. Automation is not the pressure — grazing land lost to wind farms, mining and forestry is, along with winters that swing between thaw and ice crust and destroy natural forage. Most herding households already combine herding with tourism, meat processing or wage work. That mix, rather than technology, decides who stays.
Drones are used for spotting and occasionally for nudging animals, but reindeer are semi-wild and stress badly under mechanical pressure, which costs weight and calving success. Corral separation, lassoing and handling sick animals need hands. No robotics programme is targeting a market this small in terrain this hostile.
Animal reading and corral craft remain the core, but the growing half of the job is administrative and political: subsidy and predator-compensation systems, herd data management, and participating in consultations over land use. Herders who can argue their case with documentation and grazing-history data protect their pastures better than those who only know the mountain.