■ SAFE RISK ■ Public Service & Government
Already happened, decades ago. The automation of this job is history rather than forecast, and what remains is heritage stewardship, hospitality and maintenance rather than tending a light.
“Automated lights replaced most of you. The remaining ones are tourism jobs with a view.”
Our AI replacement risk score — how we score jobs
The original job — trimming wicks, winding clockwork rotation mechanisms, standing watch, logging weather, sounding the fog signal, and keeping the optic spotless — was systematically eliminated across the twentieth century. Electrification, automatic lamp changers, solar panels with LED arrays, remote monitoring over telemetry and eventually GPS and AIS made a resident human unnecessary. Most national lighthouse authorities finished de-manning their stations by the 1990s; the last keepers were news stories precisely because they were the last.
What people do today under this title is different work. There are heritage keepers and caretakers who run visitor operations, give tours, manage bookings and holiday lets in old keepers' cottages, and maintain buildings that sit in the most corrosive environment available on Earth. There are technicians for the aids-to-navigation authority who travel between stations servicing batteries, solar arrays, optics and telemetry — a maintenance round, not a residency. And there are volunteer and trust-run stations kept alive by people who care about them.
Automation still cannot do the remaining parts. Salt-corroded ironwork, storm damage, seabird guano, failing gaskets and the physical business of getting to a rock station by boat or helicopter all need hands. Visitor operations need a person who can tell the story. But it would be dishonest to call this a resilient job in the normal sense: the risk score is low because there is almost nothing left to automate, not because the occupation is protected. The employment already shrank by an order of magnitude, and it is not coming back.
Automatability: our editorial assessment of current and near-term AI capability
The disruption is finished. Manned lighthouse keeping was automated out of existence between roughly the 1960s and the 1990s, and satellite navigation removed much of the remaining strategic case for the lights themselves. The surviving roles — heritage caretaking, visitor operations and technician rounds — are stable and small, and face no further automation pressure through 2040. There is simply not enough left to take.
Very few in the original sense. Navigation authorities automated their stations decades ago, and the lights now run on solar, LED and remote telemetry with technicians visiting on a maintenance round. The people living at lighthouses today are usually heritage caretakers, trust staff or hosts running accommodation and tours, with building maintenance as the serious part of the job.
No — ordinary automation did, long before anything resembling modern AI existed. Automatic lamp changers, electrification, remote monitoring and eventually GPS removed the need for a resident watch. It is one of the cleanest historical examples of technology eliminating an occupation, which is exactly why the surviving roles have so little automation exposure left.
Small and stable rather than secure. Positions are scarce, often seasonal or attached to a preservation trust, and depend on tourism revenue and heritage funding rather than navigation budgets. The people who last are the ones who can also do the maintenance, run the visitor operation and help keep the grant applications alive.
Practical ones. Marine and coastal building maintenance, solar and battery systems, small-boat handling and safety, plus the hospitality side — bookings, tours, interpretation and social media for the site. The combination of hands-on trade competence and visitor-facing storytelling is what makes a caretaker worth housing on a remote rock.