■ HIGH RISK ■ Construction & Trades
No robot is crawling through your attic to fish Romex past a joist any time soon. AI will eat the diagnostics, the load calculations, and the paperwork — but the person holding the strippers stays employed, with a caveat about prefab construction.
“Smart homes wire themselves. Just kidding, you're safe — nobody trusts that.”
Our AI replacement risk score — how we score jobs
An electrician's day is a mix of physical problem-solving and code compliance: pulling wire through finished walls, bending conduit, terminating panels, troubleshooting why half a kitchen went dead, and signing off that all of it satisfies the NEC. The job happens in crawlspaces, on ladders, and inside hundred-year-old houses where the previous 'electrician' was a homeowner with a death wish. Every site is different — different framing, different vintage wiring, different surprises behind the drywall — which is exactly the unstructured environment robots handle worst. A day might swing from a routine outlet swap to chasing an intermittent arc fault that only shows up when the dryer runs.
What automation is already taking: the thinking-adjacent parts. Software does load calculations, panel schedules, and code lookups that used to justify a senior guy's hourly rate. Smart panels and connected breakers self-report faults, so diagnosis increasingly starts on a phone app instead of with a multimeter. The bigger structural threat is prefabrication — modular construction moves wiring into factories, where wall panels arrive pre-wired and the on-site electrician becomes a connector of plug-and-play harnesses. That shrinks hours per house even if it never eliminates the trade.
What resists: everything involving an existing building. Retrofit, service work, EV charger installs, panel upgrades, and troubleshooting live systems require a licensed human who can improvise, get physically into awkward spaces, and legally take responsibility when something could burn a house down. Licensing and liability are a moat AI can't cross — no insurer is underwriting an unsupervised robot's panel swap. Our 55 risk score reflects a trade where demand is actually rising (electrification, EVs, solar) while the routine share of the work quietly automates.
Automatability: our editorial assessment of current and near-term AI capability
Expect serious pressure by around 2030 — not from robots with screwdrivers, but from prefab construction compressing on-site hours and smart-home diagnostics shrinking troubleshooting time. New-construction rough-in work automates fastest; service and retrofit work barely moves. Meanwhile electrification is inflating demand, so the squeeze on tasks may be masked by a boom in total work for most of the decade.
Safer than most. AI can't crawl through an attic or take legal responsibility for a panel that could start a fire, and electrification (EVs, solar, heat pumps) is driving demand up. The real pressure comes from prefab construction reducing on-site hours, not from AI directly. Our 55 risk score reflects task-level change, not job elimination.
In factories, they already do — prefab wall panels arrive pre-wired. On a job site or in an existing home, no. The environments are too varied, the spaces too tight, and the liability too serious. What changes is the ratio: more factory wiring, less field wiring, and diagnostics that start on an app.
Smart-home systems, EV charging installation, solar and battery interconnection, and connected panel diagnostics. The electricians who struggle in 2030 will be the ones who only know residential rough-in — exactly the part prefab construction is moving into factories.
Mostly helps. Code-lookup tools, load-calc software, and self-diagnosing appliances make jobs faster, which means more jobs per day and better margins for owner-operators. The hurt shows up later, if faster jobs plus prefab construction mean fewer total field hours to go around.