Will AI Replace Electricians?

What the employment figures say

The US Bureau of Labor Statistics counts 818,700 people in Electricians (SOC 47-2111) in 2024 and projects 9% growth from 2024 to 2034, much faster than the average for all occupations, with about 81,000 openings a year; it attributes that outlook to electrical demand, construction, replacement needs and renewable energy rather than AI ^{(U.S. Bureau of Labor Statistics 2025)}. That agrees with the exposure score instead of contradicting it: the task content is hard to automate and the headcount is projected to rise. One caution on the figures — the Handbook's 818,700 for 2024 comes from a different BLS product than the OEWS estimate of 712,580 employed electricians in May 2023, so the raw 106,120 difference is not a verified trend and cannot be attributed to AI ^{(U.S. Bureau of Labor Statistics 2023)}.

What the research says about electricians

Now

The present record describes a trade where hiring, not automation, is the binding constraint. Lightcast, which sells labour-market data and analytics products, reported in February 2025 that advertised wages in electrician postings had risen 18% over the prior three years and that 21% of approximately 850,000 electricians were age 55 or older, linking demand to data-center construction in central Washington ^{(Lightcast 2025)}. EC Magazine's 2026 Profile of the Electrical Contractor — an online, self-selected survey of 926 subscribers and panellists, published by a title serving the industry it surveys — found 68% of firms with steady rosters against 13% that reduced staff, while 56% reported difficulty finding trained workers and 27% reported retention problems ^{(Electrical Contractor Magazine 2026)}. Pay has moved gently rather than sharply: the BLS median annual wage was $61,590 in May 2023 and $62,350 in May 2024, a change of $760 that BLS does not attribute to AI or any other single cause ^{(U.S. Bureau of Labor Statistics 2025)}.

The next few years

Projections over the next few years point the same way as the current data. Canada's Job Bank rated the employment outlook for electricians (except industrial and power system, NOC 72200) as 'Good' for 2025-2027 in the North Shore region of British Columbia, a regional labour-market assessment with no AI attribution ^{(Government of Canada Job Bank 2025)}. The AI industry's own behaviour is consistent with that reading: Google, which sells AI products and benefits from wider AI familiarity, announced support in 2025 for the IBEW-NECA Electrical Training Alliance covering 100,000 electrical workers and 30,000 apprentices, integrating AI tools and aiming at a 70% increase in the workforce pipeline within five years ^{(Google 2025)}. That is a large AI vendor funding more electricians rather than fewer, which is not independent evaluation but is a useful signal about expected demand.

Longer term

Further out, the ceiling on exposure is set by robotics rather than by language models, and robotics lags badly. Skilled manual electrical work sits in the lowest-exposure tail across the Microsoft applicability, Eloundou and Felten measures ^{(Microsoft Research 2025)}. The most aggressive number available runs the other way: an ELECTRI International report puts automation potential for electricians at about 42%, against 50% for plumbers and carpenters and 88% for operating engineers, though it defines neither the calculation, the task list, the horizon nor the unit of measure, and comes from a body connected to the electrical contracting industry ^{(ELECTRI International 2020)}. Even read generously it describes the least automatable trade on its own list, and the duties that define the occupation — installation, testing, troubleshooting, repair, preventive maintenance, drawing interpretation and code compliance — are performed on site ^{(Statistics Canada 2021)}.

What is exposed, and what protects it

● Where the exposure is

The exposed edge of this work is the part that happens away from the conduit. IECI, an electrical-contractor association that sells education, training and member services, summarises an October 2023 report in which 507 of 676 businesses (75%) said they used AI, with use cases including productivity, back-office work and design among construction customers; it is not a representative electrician adoption rate, but it shows where the tools arrive first ^{(Independent Electrical Contractors 2023)}. Guided-assembly research also shows the manual work can be compressed: a peer-reviewed experiment on augmented-reality support for control-cabinet wiring found average wiring time fell approximately 36-37% across 48 runs, on a very small sample of 2 testers and excluding live-site shortages and errors ^{(PMC 2020)}. Faster wiring by the same licensed person is a productivity result, not a substitution result.

○ What holds

Two moats carry most of the weight. The first is physical embodiment: panels, conduit, fixtures and distribution systems sit in varied and awkward spaces, which is why manual trades dominate the zero-exposure tail of every dataset ^{(Microsoft Research 2025)}. The second is licensure and safety-critical accountability, where a qualified person signs for code compliance. The New York Electrical Inspection Agency warned practitioners on June 27, 2025 to be cautious when using the internet or AI for electrical-code research, advising them to read the NEC and consult a qualified professional ^{(New York Electrical Inspection Agency 2025)}. The agency has an interest in promoting professional code expertise, but the liability structure it points at is genuine: AI can advise, and it cannot wire a panel.

Which tasks go first

The split between machine work and human work is unusually clean here. AI already reaches the desk-bound layer — estimating, scheduling, documentation and first-pass code lookup — which is the pattern behind the small-business adoption figures reported by the electrical-contractor association ^{(Independent Electrical Contractors 2023)}. Guided-assembly tools touch the manual layer without removing it: the augmented-reality wiring study cut average time by roughly 36-37% on repetitive cabinet runs ^{(PMC 2020)}. A Boeing case report describes Google Glass with Upskill Skylight guiding technicians through aircraft wiring, about 130 miles of wire per aircraft, for a 25% reduction in wiring time with zero reported errors — though those workers are aircraft wiring technicians rather than licensed building electricians, and the account comes from a publication covering the augmented-reality market ^{(AR Insider 2021)}. What stays human is the on-site sequence itself: laying out, assembling, installing, testing, troubleshooting, repairing and maintaining wiring, fixtures, lighting, control and distribution systems, reading drawings and specifications, and ensuring code compliance ^{(Statistics Canada 2021)}. Diagnosis in a live building, where the drawing is wrong and the access is bad, is the part no current tool contests.

What to do about it

Push AI into the office side of the work, not the code side

Quoting, scheduling, job notes and customer correspondence are where the reported adoption sits among construction-adjacent businesses ^{(Independent Electrical Contractors 2023)}. Code research is the exception: the New York Electrical Inspection Agency advises reading the NEC and consulting a qualified professional rather than relying on internet or AI answers ^{(New York Electrical Inspection Agency 2025)}. Treating a model as a drafting aid and the code book as the authority keeps the liability where it belongs.

Aim your next specialism at where the load is growing

Lightcast, a vendor of labour-market data, ties rising advertised wages to data-center construction and reports that 21% of approximately 850,000 electricians are 55 or older ^{(Lightcast 2025)}. BLS names electrical demand, construction, replacement needs and renewable energy as the drivers behind about 81,000 annual openings ^{(U.S. Bureau of Labor Statistics 2025)}. Capacity work, controls and renewables sit at the intersection of both signals.

Take the funded training while it is being offered

Google's announced support for the IBEW-NECA Electrical Training Alliance targets 100,000 electrical workers and 30,000 apprentices, with AI tools folded into the curriculum ^{(Google 2025)}. Google sells AI products and benefits from broader familiarity with them, so read it as a subsidised route to fluency rather than an independent endorsement. The value to a working electrician is knowing what the tools do and do not get right before a client assumes they are infallible.

Tools and courses that fit this situation

Below are paid tools and courses that fit the situation this page describes: office-side AI fluency for people who run or will run electrical work, rather than anything that touches the trade itself. They are worth what they are worth — none of them changes the exposure picture for a licensed electrician.

Some links above are affiliate links. If you buy through them, Jobs After AI may earn a commission at no extra cost to you. It does not change which tools appear here.

Common questions

Is demand for electricians growing or shrinking?

BLS projects 9% growth in Electricians (SOC 47-2111) from 2024 to 2034 from a 2024 base of 818,700, with about 81,000 openings a year, and attributes the outlook to electrical demand, construction, replacement needs and renewable energy rather than AI ^{(U.S. Bureau of Labor Statistics 2025)}. Contractor-side reporting is consistent: 68% of firms in EC Magazine's self-selected 2026 survey had steady rosters and 56% reported difficulty finding trained workers ^{(Electrical Contractor Magazine 2026)}. Canada's Job Bank rated the North Shore, BC outlook for NOC 72200 as 'Good' for 2025-2027 ^{(Government of Canada Job Bank 2025)}.

Which parts of an electrician's job can AI already do?

The administrative layer is where use is documented — productivity, back-office work and design support among construction businesses, per a trade-body summary of an October 2023 report covering 507 of 676 businesses ^{(Independent Electrical Contractors 2023)}. Augmented-reality guidance can also speed up repetitive wiring, by roughly 36-37% in one small peer-reviewed experiment ^{(PMC 2020)}. Neither case removes the licensed person from the site.

Could robotics change this?

That is the only route by which the score would move much, and robotics currently lags language models enough that manual trades sit in the lowest-exposure tail of the Microsoft, Eloundou and Felten measures ^{(Microsoft Research 2025)}. The highest published estimate for the trade is about 42% automation potential from an industry-connected report that does not define its calculation, task list or horizon ^{(ELECTRI International 2020)}. It would take machines that handle unstructured building interiors, plus a legal route to certified sign-off, before the picture shifted.

What would have to change for this score to fall?

Three things at once: dexterous field robotics reaching varied real-world sites, a regulatory change allowing non-human code compliance, and demand drivers reversing — and BLS currently names construction, replacement needs and renewable energy as pushing demand up rather than down ^{(U.S. Bureau of Labor Statistics 2025)}. The duty list for the occupation is dominated by physical installation, testing, troubleshooting and maintenance ^{(Statistics Canada 2021)}. None of the evidence here shows movement on any of those fronts.

Is this about electricians losing jobs, or about their tasks changing?

Tasks, and mostly at the margins. The AI-Proof Score measures how much of an occupation's task content current systems can cover, which sits upstream of employment, and the official series for this job is projected to grow rather than contract ^{(U.S. Bureau of Labor Statistics 2025)}. Where tools do land, the documented effect is faster completion of the same work by the same worker ^{(PMC 2020)}.

Related occupations

The evidence behind this page

How this score was built

The AI-Proof Score is an estimate of how exposed an occupation's core tasks are to AI over roughly the next three to five years, on a 0–100 scale where higher is safer. It measures the exposure of tasks, which sits upstream of, and is softer than, job loss — it is not a probability that this job disappears, and not a claim about any individual doing it. The full derivation, the four research families behind it and its limitations are set out in the methodology, and every occupation is ranked in the AI Job Impact Report 2026.