■ HIGH RISK ■ Agriculture & Environment
Fish farming is industrializing the way poultry did — sensors, automated feeders, camera-based biomass counts — and routine husbandry labor is exactly what gets engineered out. The industry itself is growing fast, so displaced deck labor partly becomes technician work; the person who only scooped feed is the one at risk.
“Automated fish farms monitor feeding, oxygen, and growth. You monitored your phone.”
Our AI replacement risk score — how we score jobs
On a salmon pen site or a land-based farm, the daily work is husbandry at industrial scale: feeding runs, checking oxygen and temperature, inspecting nets and moorings, removing dead fish (morts), grading and moving stock between pens, delousing and vaccination operations, and endless cleaning — biofouling on nets, biofilters in recirculating systems. It's physical outdoor work on water, in weather, with living inventory that can die overnight from a plankton bloom or an oxygen crash.
This is one of the most aggressively automating corners of food production. Camera-and-AI systems now watch feeding in real time and cut off feed when pellets start falling past uninterested fish — the single biggest cost on a farm, optimized without a human on the pontoon. Computer vision counts fish, estimates biomass, and spots sea lice; underwater drones and crawling robots inspect and clean nets; automated mort collection and remote feed barges centralize what took a site crew. Land-based recirculating farms are essentially process plants, run from control rooms. The Norwegian salmon industry — the sector's Silicon Valley — treats remote, minimally-staffed sites as an explicit design goal.
Two things temper the risk. First, growth: aquaculture supplies an ever-larger share of the world's seafood, and expanding production keeps creating jobs even as labor-per-ton falls. Second, the sea fights back — storms damage gear, disease outbreaks demand rapid hands-on response, harvest and grading still need crews, and somebody must maintain all the automation in saltwater, which destroys everything. The role is migrating from farmhand to technician: fewer people scooping feed, more people flying inspection drones, tuning sensors, and making husbandry calls from dashboards. Our 55 lands on that migration.
Automatability: our editorial assessment of current and near-term AI capability
The automation is operational today — AI feeding systems, camera-based lice counting, and remote feed barges are standard on modern salmon farms, and land-based facilities open already control-room-first. Expect routine husbandry labor per ton to keep falling through around 2030 while industry growth adds technician, maintenance, and monitoring roles. Workers have this decade to move from deck labor to the control room and the toolbox.
The routine labor is — AI feeding, camera monitoring, and net-cleaning robots are already standard on advanced farms, and labor per ton keeps falling. But the industry is growing fast enough that total employment holds up better than the task list suggests, shifting toward technician, maintenance, and fish-health roles. Adaptable workers change jobs without changing industries.
Maintenance technicians (saltwater destroys equipment faster than robots can be invented), fish-health and welfare specialists whose judgment regulators increasingly require, site managers making stocking and harvest calls, and emergency-response roles for storms, disease, and oxygen events. The exposed job is routine husbandry — feeding, watching, scooping — which is precisely what the sensors took.
Genuinely yes, entered correctly. It's one of the fastest-growing food sectors, chronically short of skilled people, and modernizing so fast that technical training compounds. Enter through an aquaculture program or as a maintenance-minded hand, aim at the technology and biology sides, and our 55 risk score mostly stops applying to you — it's aimed at the unskilled-labor tier.
Cameras watch pellets fall through the pen and cut feeding the moment fish lose interest, saving the farm's biggest cost. Vision systems count lice and estimate biomass without handling fish; sensors track oxygen and temperature continuously; drones inspect nets. Humans set the parameters, respond to what the systems flag, and fix everything the ocean breaks — which is a lot.