Will AI replace Mechanical Engineers?

Not as a job, but it is already doing parts of the work. Across the 79 official task statements scored for Mechanical Engineers (United States, SOC 17-2141), 19% of the importance-weighted core work is made of tasks today's AI could already do most of. The overall exposure score is 37 out of 100 (range 31–44, band: low). That is a statement about tasks, not about headcount: this measures what AI could do, not whether any employer adopts it, whether the law allows it, or whether doing the routine parts faster creates more demand for the human parts. Figures are from release 2026-q4.1.

Which tasks in “Mechanical Engineers” can AI already do?

The highest-scoring tasks in release 2026-q4.1 are: “Write, review, or maintain engineering documentation” (81/100, very high); “Prepare or present technical or project status reports” (79/100, high); “Research computerized automotive applications, such as telemetrics, intelligent transportation systems, artificial intelligence, or automatic control” (75/100, high). Each score comes from five published 0–4 ratings turned into a number by a published formula, and each carries the model's one-sentence reason on the page.

Which tasks in “Mechanical Engineers” stay human?

About 58% of this job's task weight sits in work that scores low for AI exposure. The lowest-scoring tasks in release 2026-q4.1 are: “Prepare test stations, instrumentation, or data acquisition systems for use in specific tests of fuel cell components or systems” (0/100, minimal); “Fabricate prototypes of fuel cell components, assemblies, stacks, or systems” (0/100, minimal); “Conduct fuel cell testing projects, using fuel cell test stations, analytical instruments, or electrochemical diagnostics, such as cyclic voltammetry or impe…” (7/100, minimal). Low scores usually mean the task needs a body in a room, a legally accountable human, or trust built in real time. Those are the three things the scoring rubric treats as gates rather than obstacles.

What should someone working in “Mechanical Engineers” do about AI?

Start from the ledger rather than the headline: 19% of this job's weighted core work is exposed, and roughly 58% is not. The practical move is to spend more of your week on the tasks that score low, the ones above, and to get fluent at directing AI through the tasks that score high, because those are the parts that change whether or not you are ready for them. This page does not predict your job, and nothing here is career advice tailored to you: the score describes the occupation, not the person.

How is the AI exposure score for Mechanical Engineers calculated?

Each official task statement for the occupation is rated on five published 0–4 dimensions (output replicability, physical embodiment, licensed accountability, real-time human trust, and data availability) by claude-opus-5 using scoring prompt task_scoring_v1.0. The model never writes the score; a published formula turns the five ratings into a 0–100 number, so every score can be recomputed by hand. The occupation figure is the importance-weighted mean across 79 scored tasks. The prompt, the rubric, the formula and the full dataset are published at https://futureproof.collab365.com/method and https://futureproof.collab365.com/data/2026-q4.1 under CC BY 4.0.

US dataswitch to UK

Mechanical Engineers

reading and interpreting blueprints, technical drawings, schematics or computer-generated reports, reading current literature and validating design of fuel cells. If that's your week, this page is about your job.

The honest answer

AI is already taking a real slice of the routine work here: writing, reviewing or maintaining engineering documentation. That is a slice of tasks, not of you.

Your move: what you can actually do about this ↓

Where the work sits, by task weight

shifting to AI: 19%
changing shape: 23%
staying human: 58%

These bars are tasks changing hands, not people being counted out. The ledger below shows which.

Whole-job exposure score 37 out of 100 (31–44 allowing for uncertainty): low exposure, across 79 scored tasks. The number is the support for the sentence above it, not a headline about anyone’s future.

How we know this

What is measured: Every published task statement for mechanical engineers is rated on five dimensions: can a model produce the output, does the work need a body in a room, does it need a legally accountable person, does it depend on a person being trusted in the moment, and how much data exists. A published formula turns those five ratings into the score; the model never writes the number.

How the bar is built: Each task’s share of the bar is its published importance weight, so a task you do all day counts for more than one you do twice a year.

Release: 2026-q4.1, scores computed 2026-08-04. Read the full method.

That slice is not coming back; the core of the job, conducting or directing system-level automotive testing, stays yours. The tools change hands, the accountability doesn't.

Your week, as this page understands it

Perform engineering duties in planning and designing tools, engines, machines, and other mechanically functioning equipment. Oversee installation, operation, maintenance, and repair of equipment such as centralized heat, gas, water, and steam systems. The job title says “mechanical engineers”. The real job is the part underneath: conducting or directing system-level automotive testing. That is the thing someone has to be right about.

The exposed part of this job is specific, and we won’t pretend it is coming back. But mechanical engineers is not one task. It is 79 scored ones, and they are nowhere near equally exposed. The most durable of them, on this evidence, is conducting or directing system-level automotive testing, and the ledger below shows exactly why.

Where the work sits, by task weight

shifting to AI: 19%
changing shape: 23%
staying human: 58%

These bars are tasks changing hands, not people being counted out. The ledger below shows which.

How we know this

How the bar is built: Each task’s share of the bar is its published importance weight, so a task you do all day counts for more than one you do twice a year.

Release: 2026-q4.1, scores computed 2026-08-04. Read the full method.

Your job, task by task

These are the official task statements for this occupation, in plain English, sorted by what the evidence says is happening to each one. The official wording sits under every line so you can check the rewrite against it.

8 of this occupation's scored task statements carry a score that was measured under a different occupation's context, because the statement is shared between jobs and has only been scored once so far. Each one names the occupation it was measured under in the free routes below; none is presented as a measurement for this job.
Task clusters are not derived in this release, so the task-cluster field is empty and no Collab365 Space routing is attached to this occupation yet.

Shifting to AI

14 tasks

Tasks today’s tools can already do most of. This is the part we will not soften: where these rows are the bulk of your week, the week changes.

Reading and interpreting blueprints, technical drawings, schematics or computer-generated reports
It is the same call made over and over on blueprints, technical drawings, schematics or computer-generated reports, with a right answer to check it against. That is what a model is trained on.
importance 4 · Core
Source: “Read and interpret blueprints, technical drawings, schematics, or computer-generated reports.” (O*NET task statement)
How this row was scored
Exposure score: 65 out of 100 (58–72 allowing for uncertainty): high exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: AI reads technical reports very well and can interpret many drawings, though complex engineering drawings still trip it up.
The five ratings: output a model can produce 2/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 4/4.
Analyzing fuel cell or related test data
This is reading one thing and writing another: fuel cell in, a record out. That is the shape today's tools are built for.
importance 4 · Core
Source: “Analyze fuel cell or related test data, using statistical software.” (O*NET task statement)
How this row was scored
Exposure score: 66 out of 100 (62–70 allowing for uncertainty): high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Statistical analysis of test data is a standard, well-documented job that AI produces to a usable standard.
The five ratings: output a model can produce 3/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Writing, reviewing or maintaining engineering documentation
This is reading one thing and writing another: documentation in, a record out. That is the shape today's tools are built for.
importance 4 · Core
Source: “Write, review, or maintain engineering documentation.” (O*NET task statement)
How this row was scored
Exposure score: 81 out of 100 (77–85 allowing for uncertainty): very high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Writing and keeping engineering documents up to date is document work AI does at least as well as a practitioner.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Preparing or presenting technical or project status reports
This is reading one thing and writing another: technical or project status reports in, a record out. That is the shape today's tools are built for.
importance 4 · Core
Source: “Prepare or present technical or project status reports.” (O*NET task statement)
How this row was scored
Exposure score: 79 out of 100 (75–83 allowing for uncertainty): high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Pulling project data into a clear status report is exactly what AI does reliably; presenting it still involves a person.
The five ratings: output a model can produce 4/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 1/4 · how much data exists 3/4.

Changing shape

19 tasks

Tasks where the machine takes the producing and a person keeps the checking, the deciding, or the answering-for-it. For most jobs this is the biggest group, and it is where "transformation, not termination" is literally visible.

Performing failure, variation or root cause analyses
The software now makes the first pass at failure, variation or root cause analyses, but the deciding part still needs a person. So the job becomes checking and deciding rather than producing.
importance 4 · Core
Source: “Perform failure, variation, or root cause analyses.” (O*NET task statement)
How this row was scored
Exposure score: 49 out of 100 (42–56 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Root cause analysis follows documented methods on data, which AI works through well, though inspecting parts may be needed.
The five ratings: output a model can produce 3/4 · needs a body in a room 1/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Defining specifications for fuel cell materials
The software now makes the first pass at specifications, but the deciding part still needs a person. So the job becomes checking and deciding rather than producing.
importance 4 · Core
Source: “Define specifications for fuel cell materials.” (O*NET task statement)
How this row was scored
Exposure score: 44 out of 100 (37–51 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: mistakes that are cheap to catch.
The rating behind it: AI drafts material specifications from published data, though the numbers that matter come from a company's own testing.
The five ratings: output a model can produce 2/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 2/4.
Reading current literature
The software now makes the first pass at current literature, but the deciding part still needs a person. So the job becomes checking and deciding rather than producing.
importance 4 · Core
Source: “Read current literature, attend meetings or conferences, or talk with colleagues to stay abreast of new technology or competitive products.” (O*NET task statement)
How this row was scored
Exposure score: 48 out of 100 (41–55 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: AI is very good at digesting new research, though conferences and conversations with peers still deliver things papers do not.
The five ratings: output a model can produce 3/4 · needs a body in a room 1/4 · needs an accountable person 0/4 · needs to be trusted in the moment 1/4 · how much data exists 3/4.

Staying human

46 tasks

Tasks that stay with a person, because they happen in the physical world, because the rules need someone accountable, or because the value is that a specific person does them.

Conducting or directing system-level automotive testing
This work happens in the physical world: system-level automotive, in a real place. Software cannot follow it there.
importance 4 · Core
Source: “Conduct or direct system-level automotive testing.” (O*NET task statement)
How this row was scored
Exposure score: 7 out of 100 (0–14 allowing for uncertainty): minimal exposure, medium confidence.
Why it sits in this group: work that happens in the physical world.
The rating behind it: Test plans and results analysis can be automated, but vehicle testing happens on tracks and rigs.
The five ratings: output a model can produce 1/4 · needs a body in a room 3/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 2/4.
Providing technical direction to other engineers or engineering support personnel
The value here is that a specific person handles technical direction and stands behind it. That is earned, not computed.
importance 4 · Core
Source: “Provide technical direction to other engineers or engineering support personnel.” (O*NET task statement)
How this row was scored
Exposure score: 21 out of 100 (14–28 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the value is that a specific person does it.
The rating behind it: AI supplies technical answers, but guiding a team of engineers depends on judgement and trust built with those people.
The five ratings: output a model can produce 2/4 · needs a body in a room 1/4 · needs an accountable person 0/4 · needs to be trusted in the moment 3/4 · how much data exists 2/4.
Planning or conducting experiments to validate new materials
This work happens in the physical world: experiments, in a real place. Software cannot follow it there.
importance 4 · Core
Source: “Plan or conduct experiments to validate new materials, optimize startup protocols, reduce conditioning time, or examine contaminant tolerance.” (O*NET task statement)
How this row was scored
Exposure score: 22 out of 100 (15–29 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: mistakes that are cheap to catch; work that happens in the physical world.
The rating behind it: Designing an experiment is something AI helps with, but running it means handling materials and test hardware in a lab.
The five ratings: output a model can produce 2/4 · needs a body in a room 2/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 2/4.

Show the other 69 tasks

Researching computerized automotive applications, such as telemetrics, intelligent transportation systems, artificial intelligence or automatic control
shifting to AI
This is reading one thing and writing another: computerized automotive applications in, a record out. That is the shape today's tools are built for.
importance 3 · Core
Source: “Research computerized automotive applications, such as telemetrics, intelligent transportation systems, artificial intelligence, or automatic control.” (O*NET task statement)
How this row was scored
Exposure score: 75 out of 100 (71–79 allowing for uncertainty): high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Surveying what is possible in connected and automated vehicle technology is literature research AI does quickly and well.
The five ratings: output a model can produce 3/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Researching and analyzing customer design proposals
shifting to AI
This is reading one thing and writing another: customer design proposals in, a record out. That is the shape today's tools are built for.
importance 4 · Core
Source: “Research and analyze customer design proposals, specifications, manuals, or other data to evaluate the feasibility, cost, or maintenance requirements of designs or applications.” (O*NET task statement)
How this row was scored
Exposure score: 66 out of 100 (62–70 allowing for uncertainty): high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Reading proposals and specifications and working out cost and feasibility is document analysis AI produces to a usable standard.
The five ratings: output a model can produce 3/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Writing performance requirements for product development or engineering projects
shifting to AI
This is reading one thing and writing another: performance requirements in, a record out. That is the shape today's tools are built for.
importance 3 · Core
Source: “Write performance requirements for product development or engineering projects.” (O*NET task statement)
How this row was scored
Exposure score: 66 out of 100 (62–70 allowing for uncertainty): high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Drafting a requirements document from project goals is exactly the kind of structured writing AI produces well.
The five ratings: output a model can produce 3/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Writing technical reports or proposals related to engineering projects
shifting to AI
This is reading one thing and writing another: technical reports in, a record out. That is the shape today's tools are built for.
importance 4 · Core
Source: “Write technical reports or proposals related to engineering projects.” (O*NET task statement)
How this row was scored
Exposure score: 66 out of 100 (62–70 allowing for uncertainty): high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Technical report and proposal writing is a strong point for AI, with an engineer checking the numbers and claims.
The five ratings: output a model can produce 3/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Calculate the efficiency or power output of a fuel cell system or process
shifting to AI
This is reading one thing and writing another: the efficiency in, a record out. That is the shape today's tools are built for.
importance 3 · Core
Source: “Calculate the efficiency or power output of a fuel cell system or process.” (O*NET task statement)
How this row was scored
Exposure score: 66 out of 100 (62–70 allowing for uncertainty): high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Efficiency and power output calculations follow published formulas applied to system data.
The five ratings: output a model can produce 3/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Developing or integrating control feature requirements
shifting to AI
This is reading one thing and writing another: control feature requirements in, a record out. That is the shape today's tools are built for.
importance 3 · Core
Source: “Develop or integrate control feature requirements.” (O*NET task statement)
How this row was scored
Exposure score: 66 out of 100 (59–73 allowing for uncertainty): high exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Turning goals into written control requirements is structured technical writing AI produces to a usable standard.
The five ratings: output a model can produce 3/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Developing engineering specifications or cost estimates for automotive design concepts
shifting to AI
This is reading one thing and writing another: specifications in, a record out. That is the shape today's tools are built for.
importance 4 · Core
Source: “Develop engineering specifications or cost estimates for automotive design concepts.” (O*NET task statement)
How this row was scored
Exposure score: 66 out of 100 (62–70 allowing for uncertainty): high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Specifications and cost estimates draw on documented data and templates, which AI turns into usable drafts.
The five ratings: output a model can produce 3/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Building models for algorithm or controlling feature verification testing
shifting to AI
This is reading one thing and writing another: models in, a record out. That is the shape today's tools are built for.
importance 4 · Core
Source: “Build models for algorithm or control feature verification testing.” (O*NET task statement)
How this row was scored
Exposure score: 66 out of 100 (59–73 allowing for uncertainty): high exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Building simulation models to check control software is coding work AI is genuinely strong at.
The five ratings: output a model can produce 3/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Evaluating the power output, system cost or environmental impact of new hydrogen or non-hydrogen fuel cell system designs
shifting to AI
This is reading one thing and writing another: the power output in, a record out. That is the shape today's tools are built for.
importance 3 · Core
Source: “Evaluate the power output, system cost, or environmental impact of new hydrogen or non-hydrogen fuel cell system designs.” (O*NET task statement)
How this row was scored
Exposure score: 66 out of 100 (59–73 allowing for uncertainty): high exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Comparing designs on output, cost and environmental impact is a modelling job AI produces to a usable standard.
The five ratings: output a model can produce 3/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Providing feedback to design engineers on customer problems or needs
shifting to AI
This is reading one thing and writing another: feedback in, a record out. That is the shape today's tools are built for.
importance 3 · Core
Source: “Provide feedback to design engineers on customer problems or needs.” (O*NET task statement)
How this row was scored
Exposure score: 64 out of 100 (60–68 allowing for uncertainty): high exposure, high confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Summarising what customers are struggling with and passing it to designers is written work AI does to a usable standard.
The five ratings: output a model can produce 3/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 1/4 · how much data exists 3/4.
Recommending the use of utility or energy services that minimize carbon footprints
changing shape
The software now makes the first pass at the use of utility, but the deciding part still needs a person. So the job becomes checking and deciding rather than producing.
importance 4 · Supplemental
Source: “Recommend the use of utility or energy services that minimize carbon footprints.” (O*NET task statement)
How this row was scored
Exposure score: 56 out of 100 (49–63 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Comparing energy options and their carbon impact is research and analysis AI handles well, with an engineer confirming it.
The five ratings: output a model can produce 3/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 1/4 · how much data exists 3/4.
Estimating costs or submitting bids
changing shape
The software now makes the first pass at costs, but the deciding part still needs a person. So the job becomes checking and deciding rather than producing.
importance 3 · Core
Source: “Estimate costs or submit bids for engineering, construction, or extraction projects.” (O*NET task statement)
How this row was scored
Exposure score: 56 out of 100 (49–63 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Cost models and past bids give AI plenty to work from, though pricing a job to win it is a judgement call.
The five ratings: output a model can produce 3/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 1/4 · how much data exists 3/4.
Coordinating fuel cell engineering or testing schedules with departments outside engineering
changing shape
The software now makes the first pass at fuel cell, but the part that matters is a person saying it and standing behind it. So the job becomes checking and deciding rather than producing.
importance 3 · Core
Source: “Coordinate fuel cell engineering or test schedules with departments outside engineering, such as manufacturing.” (O*NET task statement)
How this row was scored
Exposure score: 53 out of 100 (46–60 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; the value is that a specific person does it.
The rating behind it: Scheduling and chasing across teams is well suited to software, though the negotiating between departments involves people.
The five ratings: output a model can produce 3/4 · needs a body in a room 0/4 · needs an accountable person 0/4 · needs to be trusted in the moment 2/4 · how much data exists 3/4.
Simulating or modeling fuel cell
changing shape
The software now makes the first pass at fuel cell, but the deciding part still needs a person. So the job becomes checking and deciding rather than producing.
importance 3 · Core
Source: “Simulate or model fuel cell, motor, or other system information, using simulation software programs.” (O*NET task statement)
How this row was scored
Exposure score: 50 out of 100 (43–57 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: AI can set up and interpret many simulations, though specialist modelling tools still need an experienced engineer driving them.
The five ratings: output a model can produce 2/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Creating design alternatives for vehicle components
changing shape
The software now makes the first pass at design alternatives, but the deciding part still needs a person. So the job becomes checking and deciding rather than producing.
importance 4 · Core
Source: “Create design alternatives for vehicle components, such as camless or dual-clutch engines or alternative air-conditioning systems, to increase fuel efficiency.” (O*NET task statement)
How this row was scored
Exposure score: 50 out of 100 (43–57 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Generating and comparing design options on paper is something AI does quickly, though engineers pick and prove the winner.
The five ratings: output a model can produce 2/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Developing specifications for vehicles powered by alternative fuels or alternative power methods
changing shape
The software now makes the first pass at specifications, but the deciding part still needs a person. So the job becomes checking and deciding rather than producing.
importance 4 · Core
Source: “Develop specifications for vehicles powered by alternative fuels or alternative power methods.” (O*NET task statement)
How this row was scored
Exposure score: 50 out of 100 (43–57 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Specification writing from published performance data is well within what AI produces, subject to engineering review.
The five ratings: output a model can produce 2/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Establishing or coordinating the maintenance or safety procedures
changing shape
The software now makes the first pass at the maintenance, but the deciding part still needs a person. So the job becomes checking and deciding rather than producing.
importance 3 · Supplemental
Source: “Establish or coordinate the maintenance or safety procedures, service schedule, or supply of materials required to maintain machines or equipment in the prescribed condition.” (O*NET task statement)
How this row was scored
Exposure score: 49 out of 100 (42–56 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Maintenance schedules and safety procedures follow documented patterns, so AI produces usable drafts an engineer then checks.
The five ratings: output a model can produce 3/4 · needs a body in a room 1/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Developing or implementing operating methods or procedures
changing shape
The software now makes the first pass at methods, but the deciding part still needs a person. So the job becomes checking and deciding rather than producing.
importance 4 · Core
Source: “Develop or implement operating methods or procedures.” (O*NET task statement)
How this row was scored
Exposure score: 49 out of 100 (42–56 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Procedure documents follow familiar patterns, so AI produces usable drafts, though putting them into practice involves the workplace.
The five ratings: output a model can produce 3/4 · needs a body in a room 1/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Evaluating mechanical designs or prototypes for energy performance or environmental impact
changing shape
The software now makes the first pass at mechanical designs, but the deciding part still needs a person. So the job becomes checking and deciding rather than producing.
importance 3 · Supplemental
Source: “Evaluate mechanical designs or prototypes for energy performance or environmental impact.” (O*NET task statement)
How this row was scored
Exposure score: 49 out of 100 (42–56 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Modelling how much energy a design uses is well-documented calculation AI does reliably, though prototype checks happen in the workshop.
The five ratings: output a model can produce 3/4 · needs a body in a room 1/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Reading current literature
changing shape
The software now makes the first pass at current literature, but the deciding part still needs a person. So the job becomes checking and deciding rather than producing.
importance 3 · Core
Source: “Read current literature, attend meetings or conferences, or talk with colleagues to stay abreast of new automotive technology or competitive products.” (O*NET task statement)
How this row was scored
Exposure score: 48 out of 100 (41–55 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: AI summarises new technical literature very well, though meeting peers at events surfaces things not yet written down.
The five ratings: output a model can produce 3/4 · needs a body in a room 1/4 · needs an accountable person 0/4 · needs to be trusted in the moment 1/4 · how much data exists 3/4.
Identifying or defining vehicle and system integration challenges for fuel cell vehicles
changing shape
The software now makes the first pass at vehicle, but the deciding part still needs a person. So the job becomes checking and deciding rather than producing.
importance 3 · Core
Source: “Identify or define vehicle and system integration challenges for fuel cell vehicles.” (O*NET task statement)
How this row was scored
Exposure score: 44 out of 100 (37–51 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: mistakes that are cheap to catch.
The rating behind it: Listing the likely integration problems draws on published engineering knowledge that AI organises well.
The five ratings: output a model can produce 2/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 2/4.
Establishing production or quality control standards
changing shape
The software now makes the first pass at production, but someone has to be answerable for the result, and it cannot be the software. So the job becomes checking and deciding rather than producing.
importance 4 · Core
Source: “Establish production or quality control standards.” (O*NET task statement)
How this row was scored
Exposure score: 43 out of 100 (36–50 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; someone qualified has to answer for it.
The rating behind it: Standards documents follow published patterns AI drafts well, though a qualified engineer decides what the limits should be.
The five ratings: output a model can produce 2/4 · needs a body in a room 0/4 · needs an accountable person 2/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Designing control systems or algorithms
changing shape
The software now makes the first pass at control systems, but someone has to be answerable for the result, and it cannot be the software. So the job becomes checking and deciding rather than producing.
importance 4 · Core
Source: “Design control systems or algorithms for purposes such as automotive energy management, emissions management, or increased operational safety or performance.” (O*NET task statement)
How this row was scored
Exposure score: 43 out of 100 (36–50 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; someone qualified has to answer for it.
The rating behind it: AI writes and refines control code well, but safety and emissions systems need a responsible engineer behind them.
The five ratings: output a model can produce 2/4 · needs a body in a room 0/4 · needs an accountable person 2/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Designing vehicles for increased recyclability or use of natural
changing shape
The software now makes the first pass at vehicles, but someone has to be answerable for the result, and it cannot be the software. So the job becomes checking and deciding rather than producing.
importance 3 · Core
Source: “Design vehicles for increased recyclability or use of natural, renewable, or recycled materials in vehicle construction.” (O*NET task statement)
How this row was scored
Exposure score: 43 out of 100 (36–50 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; someone qualified has to answer for it.
The rating behind it: Recyclable material options are documented and comparable by software, though the vehicle design itself needs qualified engineers.
The five ratings: output a model can produce 2/4 · needs a body in a room 0/4 · needs an accountable person 2/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Assisting drafters in developing the structural design of products
changing shape
The software now makes the first pass at drafters, but the deciding part still needs a person. So the job becomes checking and deciding rather than producing.
importance 3 · Core
Source: “Assist drafters in developing the structural design of products, using drafting tools or computer-assisted drafting equipment or software.” (O*NET task statement)
How this row was scored
Exposure score: 43 out of 100 (36–50 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: AI helps generate and check drawings, though turning a rough idea into a correct structural design still needs an engineer.
The five ratings: output a model can produce 2/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 1/4 · how much data exists 3/4.
Studying industrial processes to maximize the efficiency of equipment applications
changing shape
The software now makes the first pass at industrial processes, but the deciding part still needs a person. So the job becomes checking and deciding rather than producing.
importance 3 · Supplemental
Source: “Study industrial processes to maximize the efficiency of equipment applications, including equipment placement.” (O*NET task statement)
How this row was scored
Exposure score: 43 out of 100 (36–50 allowing for uncertainty): partial exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Process data and layout options can be analysed by software, though seeing how a plant really runs adds context.
The five ratings: output a model can produce 2/4 · needs a body in a room 1/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Coordinating production activities with other functional units
staying human
The value here is that a specific person handles production activities and stands behind it. That is earned, not computed.
importance 4 · Core
Source: “Coordinate production activities with other functional units, such as procurement, maintenance, or quality control.” (O*NET task statement)
How this row was scored
Exposure score: 39 out of 100 (32–46 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; the value is that a specific person does it.
The rating behind it: Scheduling and status chasing across departments is well handled by software, though the negotiating is done by people.
The five ratings: output a model can produce 3/4 · needs a body in a room 1/4 · needs an accountable person 0/4 · needs to be trusted in the moment 2/4 · how much data exists 3/4.
Providing technical customer service
staying human
The value here is that a specific person handles technical customer service and stands behind it. That is earned, not computed.
importance 3 · Core
Source: “Provide technical customer service.” (O*NET task statement)
How this row was scored
Exposure score: 39 out of 100 (32–46 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: reading one thing and writing another; the same decision, made over and over; the value is that a specific person does it.
The rating behind it: AI answers technical product questions well, though customers with a serious problem often want a named engineer.
The five ratings: output a model can produce 3/4 · needs a body in a room 1/4 · needs an accountable person 0/4 · needs to be trusted in the moment 2/4 · how much data exists 3/4.
Recommending design modifications to eliminate machine or system malfunctions
staying human
The ratings behind this row put design modifications well outside what today's tools can do on their own.
importance 4 · Core
Source: “Recommend design modifications to eliminate machine or system malfunctions.” (O*NET task statement)
How this row was scored
Exposure score: 38 out of 100 (31–45 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Suggesting a design change to stop a recurring fault is analysis AI does well, though an engineer confirms it works.
The five ratings: output a model can produce 2/4 · needs a body in a room 1/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Designing test control apparatus or equipment or developing procedures for testing products
staying human
The ratings behind this row put test control apparatus well outside what today's tools can do on their own.
importance 3 · Supplemental
Source: “Design test control apparatus or equipment or develop procedures for testing products.” (O*NET task statement)
How this row was scored
Exposure score: 38 out of 100 (31–45 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Writing test procedures is documented work AI drafts well; designing the rig that carries them out needs engineering judgement.
The five ratings: output a model can produce 2/4 · needs a body in a room 1/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Developing calibration methodologies, test methodologies or tools
staying human
The ratings behind this row put calibration methodologies, test methodologies or tools well outside what today's tools can do on their own.
importance 4 · Core
Source: “Develop calibration methodologies, test methodologies, or tools.” (O*NET task statement)
How this row was scored
Exposure score: 38 out of 100 (31–45 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Writing up a test or calibration method is structured documentation AI drafts well, though it comes from hands-on experience.
The five ratings: output a model can produce 2/4 · needs a body in a room 1/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Altering or modifying designs to obtain specified functional or operational performance
staying human
The ratings behind this row put designs well outside what today's tools can do on their own.
importance 4 · Core
Source: “Alter or modify designs to obtain specified functional or operational performance.” (O*NET task statement)
How this row was scored
Exposure score: 38 out of 100 (31–45 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Design tweaks to hit a performance target can be modelled by software, with an engineer confirming the change works.
The five ratings: output a model can produce 2/4 · needs a body in a room 1/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Developing or testing models of alternate designs or processing methods to assess feasibility
staying human
The ratings behind this row put models of alternate designs well outside what today's tools can do on their own.
importance 3 · Core
Source: “Develop or test models of alternate designs or processing methods to assess feasibility, sustainability, operating condition effects, potential new applications, or necessity of modification.” (O*NET task statement)
How this row was scored
Exposure score: 38 out of 100 (31–45 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch.
The rating behind it: Building and comparing design models on a computer is work AI supports well, though physical testing and judgement still count.
The five ratings: output a model can produce 2/4 · needs a body in a room 1/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Managing fuel cell battery hybrid system architecture
staying human
The ratings behind this row put fuel cell battery hybrid system architecture well outside what today's tools can do on their own.
importance 4 · Core
Source: “Manage fuel cell battery hybrid system architecture, including sizing of components, such as fuel cells, energy storage units, or electric drives.” (O*NET task statement)
How this row was scored
Exposure score: 37 out of 100 (30–44 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: mistakes that are cheap to catch.
The rating behind it: Sizing calculations are the kind of modelling software does well, though the architecture choices carry engineering responsibility.
The five ratings: output a model can produce 2/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 1/4 · how much data exists 2/4.
Conducting automotive design reviews
staying human
The value here is that a specific person handles automotive design reviews and stands behind it. That is earned, not computed.
importance 4 · Core
Source: “Conduct automotive design reviews.” (O*NET task statement)
How this row was scored
Exposure score: 35 out of 100 (28–42 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; the value is that a specific person does it.
The rating behind it: AI can pre-check a design against requirements, but a review meeting is a group of engineers challenging each other.
The five ratings: output a model can produce 2/4 · needs a body in a room 0/4 · needs an accountable person 1/4 · needs to be trusted in the moment 2/4 · how much data exists 3/4.
Recommending or implementing changes to fuel cell system designs
staying human
The ratings behind this row put changes well outside what today's tools can do on their own.
importance 4 · Core
Source: “Recommend or implement changes to fuel cell system designs.” (O*NET task statement)
How this row was scored
Exposure score: 33 out of 100 (26–40 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: mistakes that are cheap to catch.
The rating behind it: Design change recommendations rest on test results and design knowledge held inside the company.
The five ratings: output a model can produce 2/4 · needs a body in a room 1/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 2/4.
Developing or evaluating systems or methods of hydrogen storage for fuel cell applications
staying human
The ratings behind this row put systems well outside what today's tools can do on their own.
importance 3 · Supplemental
Source: “Develop or evaluate systems or methods of hydrogen storage for fuel cell applications.” (O*NET task statement)
How this row was scored
Exposure score: 33 out of 100 (26–40 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: mistakes that are cheap to catch.
The rating behind it: Evaluating storage options against published data suits AI, though developing a workable system involves laboratory work.
The five ratings: output a model can produce 2/4 · needs a body in a room 1/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 2/4.
Designing or analyzing automobile systems in areas
staying human
The rules require a named, qualified person to answer for automobile systems, and that person cannot be a piece of software.
importance 4 · Core
Source: “Design or analyze automobile systems in areas such as aerodynamics, alternate fuels, ergonomics, hybrid power, brakes, transmissions, steering, calibration, safety, or diagnostics.” (O*NET task statement)
How this row was scored
Exposure score: 32 out of 100 (25–39 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; someone qualified has to answer for it.
The rating behind it: Analysis of vehicle systems is well supported by AI tools, but safety-critical design carries responsibility a person must hold.
The five ratings: output a model can produce 2/4 · needs a body in a room 1/4 · needs an accountable person 2/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Designing integrated mechanical or alternative systems
staying human
The rules require a named, qualified person to answer for integrated mechanical or alternative systems, and that person cannot be a piece of software.
importance 4 · Supplemental
Source: “Design integrated mechanical or alternative systems, such as mechanical cooling systems with natural ventilation systems, to improve energy efficiency.” (O*NET task statement)
How this row was scored
Exposure score: 32 out of 100 (25–39 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; someone qualified has to answer for it.
The rating behind it: AI can propose efficient system combinations from published methods, but a qualified engineer signs off what actually gets built.
The five ratings: output a model can produce 2/4 · needs a body in a room 1/4 · needs an accountable person 2/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Specifying system components or directing modification of products to ensure conformance with engineering design
staying human
The rules require a named, qualified person to answer for system components, and that person cannot be a piece of software.
importance 4 · Core
Source: “Specify system components or direct modification of products to ensure conformance with engineering design, performance specifications, or environmental regulations.” (O*NET task statement)
How this row was scored
Exposure score: 32 out of 100 (25–39 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; someone qualified has to answer for it.
The rating behind it: Matching parts to specifications and regulations is documented work AI helps with, but a qualified engineer stands behind the choice.
The five ratings: output a model can produce 2/4 · needs a body in a room 1/4 · needs an accountable person 2/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Conferring with engineers or other personnel to implement operating procedures
staying human
The value here is that a specific person handles engineers and stands behind it. That is earned, not computed.
importance 4 · Core
Source: “Confer with engineers or other personnel to implement operating procedures, resolve system malfunctions, or provide technical information.” (O*NET task statement)
How this row was scored
Exposure score: 30 out of 100 (23–37 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; the value is that a specific person does it.
The rating behind it: AI can prepare the technical answers, but sorting out a live problem with colleagues is a conversation between people.
The five ratings: output a model can produce 2/4 · needs a body in a room 1/4 · needs an accountable person 0/4 · needs to be trusted in the moment 2/4 · how much data exists 3/4.
Designing or implementing fuel cell testing or development programs
staying human
The ratings behind this row put fuel cell well outside what today's tools can do on their own.
importance 4 · Core
Source: “Design or implement fuel cell testing or development programs.” (O*NET task statement)
How this row was scored
Exposure score: 28 out of 100 (21–35 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: mistakes that are cheap to catch.
The rating behind it: Programme plans and test matrices can be drafted by AI, but running a development programme involves labs and people.
The five ratings: output a model can produce 2/4 · needs a body in a room 1/4 · needs an accountable person 1/4 · needs to be trusted in the moment 1/4 · how much data exists 2/4.
Conducting research that tests or analyzes the feasibility
staying human
This work happens in the physical world: research, in a real place. Software cannot follow it there.
importance 3 · Core
Source: “Conduct research that tests or analyzes the feasibility, design, operation, or performance of equipment, components, or systems.” (O*NET task statement)
How this row was scored
Exposure score: 25 out of 100 (18–32 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch; work that happens in the physical world.
The rating behind it: Analysis and simulation are strong points for AI, but testing real equipment means running it in a lab or plant.
The five ratings: output a model can produce 2/4 · needs a body in a room 2/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Calculating energy losses
staying human
This work happens in the physical world: energy losses, in a real place. Software cannot follow it there.
importance 4 · Supplemental
Source: “Calculate energy losses for buildings, using equipment such as computers, combustion analyzers, or pressure gauges.” (O*NET task statement)
How this row was scored
Exposure score: 25 out of 100 (18–32 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch; work that happens in the physical world.
The rating behind it: The heat-loss maths is well documented and easy for software, but taking readings with gauges happens in the building.
The five ratings: output a model can produce 2/4 · needs a body in a room 2/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Selecting or installing combined heat units
staying human
This work happens in the physical world: combined heat units, in a real place. Software cannot follow it there.
importance 3 · Supplemental
Source: “Select or install combined heat units, power units, cogeneration equipment, or trigeneration equipment that reduces energy use or pollution.” (O*NET task statement)
How this row was scored
Exposure score: 25 out of 100 (18–32 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch; work that happens in the physical world.
The rating behind it: Equipment can be selected from published data, but installing it is physical work on site.
The five ratings: output a model can produce 2/4 · needs a body in a room 2/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Researching or implementing green automotive technologies involving alternative fuels
staying human
This work happens in the physical world: green automotive technologies involving alternative fuels, in a real place. Software cannot follow it there.
importance 4 · Core
Source: “Research or implement green automotive technologies involving alternative fuels, electric or hybrid cars, or lighter or more fuel-efficient vehicles.” (O*NET task statement)
How this row was scored
Exposure score: 25 out of 100 (18–32 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch; work that happens in the physical world.
The rating behind it: The research side is well documented and suits AI, but putting new technology into vehicles is engineering and build work.
The five ratings: output a model can produce 2/4 · needs a body in a room 2/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Investigating equipment failures or difficulties to diagnose faulty operation and recommend remedial actions
staying human
This work happens in the physical world: equipment failures, in a real place. Software cannot follow it there.
importance 4 · Core
Source: “Investigate equipment failures or difficulties to diagnose faulty operation and recommend remedial actions.” (O*NET task statement)
How this row was scored
Exposure score: 25 out of 100 (18–32 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch; work that happens in the physical world.
The rating behind it: Failure data and manuals let AI suggest likely causes and fixes, though inspecting the broken equipment usually means going to it.
The five ratings: output a model can produce 2/4 · needs a body in a room 2/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Soliciting new business
staying human
The value here is that a specific person handles new business and stands behind it. That is earned, not computed.
importance 3 · Supplemental
Source: “Solicit new business.” (O*NET task statement)
How this row was scored
Exposure score: 24 out of 100 (17–31 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; the value is that a specific person does it.
The rating behind it: Outreach material writes itself these days, but winning engineering work usually rests on a relationship with the buyer.
The five ratings: output a model can produce 2/4 · needs a body in a room 1/4 · needs an accountable person 0/4 · needs to be trusted in the moment 3/4 · how much data exists 3/4.
Providing technical consultation or direction related to the development or production of fuel cell systems
staying human
The value here is that a specific person handles technical consultation and stands behind it. That is earned, not computed.
importance 4 · Core
Source: “Provide technical consultation or direction related to the development or production of fuel cell systems.” (O*NET task statement)
How this row was scored
Exposure score: 23 out of 100 (16–30 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the value is that a specific person does it.
The rating behind it: AI can pull together the technical background, though giving direction on a specialised programme rests on an expert people trust.
The five ratings: output a model can produce 2/4 · needs a body in a room 1/4 · needs an accountable person 1/4 · needs to be trusted in the moment 2/4 · how much data exists 2/4.
Designing vehicles that use lighter materials
staying human
The rules require a named, qualified person to answer for vehicles, and that person cannot be a piece of software.
importance 4 · Core
Source: “Design vehicles that use lighter materials, such as aluminum, magnesium alloy, or plastic, to improve fuel efficiency.” (O*NET task statement)
How this row was scored
Exposure score: 23 out of 100 (16–30 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; someone qualified has to answer for it.
The rating behind it: Material data is well documented, but redesigning vehicle structures for light weight is expert safety-critical engineering.
The five ratings: output a model can produce 1/4 · needs a body in a room 1/4 · needs an accountable person 2/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Planning or implementing fuel cell cost reduction or product improvement projects in collaboration with other engineers
staying human
The value here is that a specific person handles fuel cell cost reduction and stands behind it. That is earned, not computed.
importance 4 · Core
Source: “Plan or implement fuel cell cost reduction or product improvement projects in collaboration with other engineers, suppliers, support personnel, or customers.” (O*NET task statement)
How this row was scored
Exposure score: 23 out of 100 (16–30 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the value is that a specific person does it.
The rating behind it: Cost-reduction analysis suits software, though carrying a project through with suppliers and colleagues depends on working relationships.
The five ratings: output a model can produce 2/4 · needs a body in a room 1/4 · needs an accountable person 1/4 · needs to be trusted in the moment 2/4 · how much data exists 2/4.
Conducting post-service or failure analyses
staying human
This work happens in the physical world: post-service, in a real place. Software cannot follow it there.
importance 4 · Core
Source: “Conduct post-service or failure analyses, using electromechanical diagnostic principles or procedures.” (O*NET task statement)
How this row was scored
Exposure score: 22 out of 100 (15–29 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: mistakes that are cheap to catch; work that happens in the physical world.
The rating behind it: Failure patterns are documented enough for AI to suggest causes, but examining the returned hardware happens on a bench.
The five ratings: output a model can produce 2/4 · needs a body in a room 2/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 2/4.
Characterizing component or fuel cell performances by generating operating maps
staying human
This work happens in the physical world: component, in a real place. Software cannot follow it there.
importance 4 · Core
Source: “Characterize component or fuel cell performances by generating operating maps, defining operating conditions, identifying design refinements, or executing durability assessments.” (O*NET task statement)
How this row was scored
Exposure score: 22 out of 100 (15–29 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: mistakes that are cheap to catch; work that happens in the physical world.
The rating behind it: Turning test data into performance maps is analysis AI does well, but the durability testing itself runs on physical hardware.
The five ratings: output a model can produce 2/4 · needs a body in a room 2/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 2/4.
Calibrating vehicle systems, including control algorithms or other software systems
staying human
This work happens in the physical world: vehicle systems, in a real place. Software cannot follow it there.
importance 4 · Core
Source: “Calibrate vehicle systems, including control algorithms or other software systems.” (O*NET task statement)
How this row was scored
Exposure score: 22 out of 100 (15–29 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: mistakes that are cheap to catch; work that happens in the physical world.
The rating behind it: Optimising control settings from data suits software, but the final calibration is proved on a real vehicle or dynamometer.
The five ratings: output a model can produce 2/4 · needs a body in a room 2/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 2/4.
Developing, coordinating or monitoring all aspects of production
staying human
This work happens in the physical world: all aspects of production, in a real place. Software cannot follow it there.
importance 3 · Core
Source: “Develop, coordinate, or monitor all aspects of production, including selection of manufacturing methods, fabrication, or operation of product designs.” (O*NET task statement)
How this row was scored
Exposure score: 21 out of 100 (14–28 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; mistakes that are cheap to catch; work that happens in the physical world.
The rating behind it: Planning manufacturing methods is desk work AI assists with, though coordinating actual production happens on the factory floor.
The five ratings: output a model can produce 2/4 · needs a body in a room 2/4 · needs an accountable person 1/4 · needs to be trusted in the moment 1/4 · how much data exists 3/4.
Applying engineering principles or practices to emerging fields
staying human
The ratings behind this row put principles well outside what today's tools can do on their own.
importance 4 · Supplemental
Source: “Apply engineering principles or practices to emerging fields, such as robotics, waste management, or biomedical engineering.” (O*NET task statement)
How this row was scored
Exposure score: 21 out of 100 (14–28 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: the five ratings behind the score, with no single dominant reason.
The rating behind it: AI can summarise what is known in a new field, but applying engineering to genuinely novel problems needs original expert thinking.
The five ratings: output a model can produce 1/4 · needs a body in a room 1/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 2/4.
Authorizing release of fuel cell parts
staying human
The rules require a named, qualified person to answer for release of fuel cell parts, and that person cannot be a piece of software.
importance 3 · Core
Source: “Authorize release of fuel cell parts, components, or subsystems for production.” (O*NET task statement)
How this row was scored
Exposure score: 20 out of 100 (13–27 allowing for uncertainty): low exposure, medium confidence.
Why it sits in this group: someone qualified has to answer for it.
The rating behind it: The checks behind a release can be automated, but signing off parts for production is a named engineer's responsibility.
The five ratings: output a model can produce 1/4 · needs a body in a room 0/4 · needs an accountable person 3/4 · needs to be trusted in the moment 0/4 · how much data exists 2/4.
Designing fuel cell systems, subsystems, stacks, assemblies or components, such as electric traction motors or power electronics
staying human
The rules require a named, qualified person to answer for fuel cell systems, subsystems, stacks, assemblies or components, and that person cannot be a piece of software.
importance 3 · Core
Source: “Design fuel cell systems, subsystems, stacks, assemblies, or components, such as electric traction motors or power electronics.” (O*NET task statement)
How this row was scored
Exposure score: 18 out of 100 (11–25 allowing for uncertainty): minimal exposure, medium confidence.
Why it sits in this group: someone qualified has to answer for it.
The rating behind it: Original design of specialised hardware needs expert engineering; AI contributes options and checks rather than finished designs.
The five ratings: output a model can produce 1/4 · needs a body in a room 1/4 · needs an accountable person 2/4 · needs to be trusted in the moment 0/4 · how much data exists 2/4.
Researching, designing, evaluating, installing, operating or maintaining mechanical products, equipment, systems or processes to meet requirements
staying human
This work happens in the physical world: mechanical products, equipment, in a real place. Software cannot follow it there.
importance 4 · Core
Source: “Research, design, evaluate, install, operate, or maintain mechanical products, equipment, systems or processes to meet requirements.” (O*NET task statement)
How this row was scored
Exposure score: 18 out of 100 (11–25 allowing for uncertainty): minimal exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; work that happens in the physical world.
The rating behind it: The design and evaluation half is desk work AI supports, but installing, running and maintaining real equipment needs engineers on site.
The five ratings: output a model can produce 1/4 · needs a body in a room 2/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 3/4.
Integrating electric drive subsystems with other vehicle systems to optimize performance or mitigate faults
staying human
This work happens in the physical world: electric drive subsystems, in a real place. Software cannot follow it there.
importance 3 · Core
Source: “Integrate electric drive subsystems with other vehicle systems to optimize performance or mitigate faults.” (O*NET task statement)
How this row was scored
Exposure score: 14 out of 100 (7–21 allowing for uncertainty): minimal exposure, medium confidence.
Why it sits in this group: work that happens in the physical world.
The rating behind it: Control strategies can be modelled in software, but making subsystems work together in a real vehicle is hands-on engineering.
The five ratings: output a model can produce 1/4 · needs a body in a room 2/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 2/4.
Conducting research studies to develop new concepts in the field of automotive engineering
staying human
This work happens in the physical world: research studies, in a real place. Software cannot follow it there.
importance 4 · Core
Source: “Conduct research studies to develop new concepts in the field of automotive engineering.” (O*NET task statement)
How this row was scored
Exposure score: 14 out of 100 (7–21 allowing for uncertainty): minimal exposure, medium confidence.
Why it sits in this group: work that happens in the physical world.
The rating behind it: AI helps survey what exists, but developing genuinely new concepts and proving them takes original work and testing.
The five ratings: output a model can produce 1/4 · needs a body in a room 2/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 2/4.
Performing personnel functions, such as supervision of production workers, technicians, technologists or other engineers
staying human
The value here is that a specific person handles personnel functions and stands behind it. That is earned, not computed.
importance 4 · Supplemental
Source: “Perform personnel functions, such as supervision of production workers, technicians, technologists, or other engineers.” (O*NET task statement)
How this row was scored
Exposure score: 13 out of 100 (6–20 allowing for uncertainty): minimal exposure, medium confidence.
Why it sits in this group: the value is that a specific person does it.
The rating behind it: Some paperwork around managing people can be drafted, but supervising a team depends on relationships built in person.
The five ratings: output a model can produce 1/4 · needs a body in a room 1/4 · needs an accountable person 0/4 · needs to be trusted in the moment 3/4 · how much data exists 2/4.
Validating design of fuel cells
staying human
This work happens in the physical world: design of fuel cells, in a real place. Software cannot follow it there.
importance 4 · Core
Source: “Validate design of fuel cells, fuel cell components, or fuel cell systems.” (O*NET task statement)
How this row was scored
Exposure score: 12 out of 100 (5–19 allowing for uncertainty): minimal exposure, medium confidence.
Why it sits in this group: work that happens in the physical world; someone qualified has to answer for it.
The rating behind it: Validation means proving a design works in hardware, and a responsible engineer has to stand behind the result.
The five ratings: output a model can produce 1/4 · needs a body in a room 2/4 · needs an accountable person 2/4 · needs to be trusted in the moment 0/4 · how much data exists 2/4.
Developing fuel cell materials or fuel cell test equipment
staying human
This work happens in the physical world: fuel cell materials, in a real place. Software cannot follow it there.
importance 4 · Core
Source: “Develop fuel cell materials or fuel cell test equipment.” (O*NET task statement)
How this row was scored
Exposure score: 8 out of 100 (1–15 allowing for uncertainty): minimal exposure, medium confidence.
Why it sits in this group: work that happens in the physical world.
The rating behind it: AI can suggest promising material ideas, but developing them means making and testing them in a lab.
The five ratings: output a model can produce 1/4 · needs a body in a room 3/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 2/4.
Overseeing installation, operation, maintenance or repairing to ensure that machines or equipment are installed and functioning
staying human
This work happens in the physical world: installation, operation, maintenance, in a real place. Software cannot follow it there.
importance 3 · Core
Source: “Oversee installation, operation, maintenance, or repair to ensure that machines or equipment are installed and functioning according to specifications.” (O*NET task statement)
How this row was scored
Exposure score: 7 out of 100 (0–14 allowing for uncertainty): minimal exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; work that happens in the physical world.
The rating behind it: Checklists and reports can be automated, but overseeing an installation means being where the machine is being put in.
The five ratings: output a model can produce 1/4 · needs a body in a room 3/4 · needs an accountable person 1/4 · needs to be trusted in the moment 1/4 · how much data exists 3/4.
Directing the installation, operation, maintenance or repairing of renewable energy equipment
staying human
This work happens in the physical world: the installation, operation, in a real place. Software cannot follow it there.
importance 4 · Supplemental
Source: “Direct the installation, operation, maintenance, or repair of renewable energy equipment, such as heating, ventilating, and air conditioning (HVAC) or water systems.” (O*NET task statement)
How this row was scored
Exposure score: 7 out of 100 (0–14 allowing for uncertainty): minimal exposure, medium confidence.
Why it sits in this group: the same decision, made over and over; work that happens in the physical world.
The rating behind it: Plans and schedules can be produced automatically, but directing work on heating and water systems means being at the site.
The five ratings: output a model can produce 1/4 · needs a body in a room 3/4 · needs an accountable person 1/4 · needs to be trusted in the moment 1/4 · how much data exists 3/4.
Conducting fuel cell testing projects
staying human
This work happens in the physical world: fuel cell testing projects, in a real place. Software cannot follow it there.
importance 4 · Core
Source: “Conduct fuel cell testing projects, using fuel cell test stations, analytical instruments, or electrochemical diagnostics, such as cyclic voltammetry or impedance spectroscopy.” (O*NET task statement)
How this row was scored
Exposure score: 7 out of 100 (0–14 allowing for uncertainty): minimal exposure, medium confidence.
Why it sits in this group: work that happens in the physical world.
The rating behind it: Test plans and data analysis can be automated, but running cells on a test station is hands-on laboratory work.
The five ratings: output a model can produce 1/4 · needs a body in a room 3/4 · needs an accountable person 1/4 · needs to be trusted in the moment 0/4 · how much data exists 2/4.
Fabricating prototypes of fuel cell components
staying human
This work happens in the physical world: prototypes of fuel cell components, in a real place. Software cannot follow it there.
importance 4 · Core
Source: “Fabricate prototypes of fuel cell components, assemblies, stacks, or systems.” (O*NET task statement)
How this row was scored
Exposure score: 0 out of 100 (0–4 allowing for uncertainty): minimal exposure, high confidence.
Why it sits in this group: work that happens in the physical world.
The rating behind it: Making a physical prototype is workshop work with materials and tools.
The five ratings: output a model can produce 0/4 · needs a body in a room 4/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 2/4.
Preparing test stations, instrumentation or data acquisition systems for use in specific tests of fuel cell components or systems
staying human
This work happens in the physical world: test stations, instrumentation or data acquisition systems, in a real place. Software cannot follow it there.
importance 4 · Core
Source: “Prepare test stations, instrumentation, or data acquisition systems for use in specific tests of fuel cell components or systems.” (O*NET task statement)
How this row was scored
Exposure score: 0 out of 100 (0–4 allowing for uncertainty): minimal exposure, high confidence.
Why it sits in this group: work that happens in the physical world.
The rating behind it: Setting up a test rig and wiring in the instruments is physical laboratory work.
The five ratings: output a model can produce 0/4 · needs a body in a room 4/4 · needs an accountable person 0/4 · needs to be trusted in the moment 0/4 · how much data exists 2/4.

What this job pays, and how many people do it

Median pay: $104,110a year, the middle of the range, so half earn more and half earn less.bls-oews, 2025 · May 2025 estimates (national_M2025_dl.xlsx)
How we know this
Source: bls-oews
Reference period: May 2025 estimates (national_M2025_dl.xlsx)
Rounding: Shown as published.
People doing this job: 296,810in the US, 2025.bls-oews · May 2025 estimates (national_M2025_dl.xlsx)

What is deliberately not here: a forecast of how many of these jobs exist in ten years. Where an official projection exists for a market we publish it with its vintage; where it does not, we leave the space empty rather than borrow the other country’s number.

Why this is shifting

The reason is boringly specific. Most of what is shifting here is reading one thing and writing another: blueprints, technical drawings, schematics or computer-generated reports in, a record out. The rows above are exactly that shape: writing, reviewing or maintaining engineering documentation and reading and interpreting blueprints, technical drawings, schematics or computer-generated reports. What it cannot do is be there in the room, and that is still where system-level automotive gets done. Which is why this page talks about your tasks changing, not your job ending.

Your move

Over a pint: what I’d tell you if you were my friend

Start with what does not change: conducting or directing system-level automotive testing is the middle of this job, and the evidence on this page says it stays with a person.

So, given all that: 19% of this job's task weight sits in rows the software is already learning, 23% in rows that change shape rather than disappear, and 58% in rows it is nowhere near. That is the position, measured across 79 scored tasks. It is not a forecast about you.

So the thing worth your attention is not the job going away. It is the layer around it. Writing, reviewing or maintaining engineering documentation is the part turning into software, and being the person who understands that layer is worth money.

This week: one thing

Ask the one question. Find whoever is bringing new software into your workplace (the manager, the office, whoever runs the system) and ask them what it is meant to do to blueprints, technical drawings, schematics or computer-generated reports, and what it is not meant to touch. Ten minutes, this week, before anyone decides it for you.

What you end up holding: a straight answer about what is actually being rolled out, and when
How long it takes: ten minutes

If there’s nobody obvious to ask, or you’d rather not ask your manager: Put the same question to your union rep, your shift lead or the person who has been there longest, in person, over a break. Same ten minutes, same answer, and you will usually get a straighter one. Write down what they say. The note is the artifact, and it tells you whether system-level automotive is in scope or not. Nothing to log into, no license needed.

Over the next 90 days

Get inside the tool rollout rather than waiting for it. Over the next ninety days, ask to be in the group that tests, checks or signs off whatever new system arrives near performing failure, variation or root cause analyses. It is usually an unglamorous seat that nobody fights for, and it is the one that decides how the software is used on your job rather than to it.

Over the next 12 months

On this evidence I would not retrain out of this job, and I will say that plainly rather than hedge it. The task list here is dominated by work that stays with a person. What I would do with a year is get formally recognised for the layer around it (the systems, the compliance, the planning), so you are the one who understands the software instead of the one it is done to. Before you pay for anything, use CareerOneStop - Find local training. It is free, it is the Labor Department's own service, and it is listed below with the rest of the free routes.

The roads out of here, and why I am not sending you down them

I looked at the obvious moves out of this job, and here is what I found.

I checked the 12 nearest US occupations to mechanical engineers (nearest by the work that AI is not taking, not by job title), and none of them survived. The closest was mechanical engineering technologists and technicians: only about 6% of its durable work is work you already do, it pays 28.4% less and there are far fewer of those jobs than of yours. And on the numbers you do not need one. This job scores 37/100 here, with only 19% of the task list in the top band, and “conduct or direct system-level automotive testing” is not work that hands over cleanly. None of them beats deepening what you already have.

How that was checked: this job was compared against all 830 US occupations in this release on their official task statements, and the 12 nearest were examined one by one. A move that turns on an industry, an employer or a qualification rather than on the work itself will not show up in a check like that. And this release carries no licence register, so anything you are weighing needs that looked up separately.

3 moves I checked and rejected

These are the obvious-looking jumps. They are here with their reasons rather than quietly dropped, because the ones that fail are worth knowing about. It is one less thing to turn over at night.

Mechanical Engineering Technologists and Technicians
Why it looked obvious: It came up as a near neighbour on the work AI is not taking: you already design vehicles that use lighter materials, and their equivalent is to improve fuel efficiency by testing vehicles or components that use lighter materials. Across both published task lists that is about 6% of the durable work in that job.
Why I am not recommending it: Almost none of it is work you already do: about 6% of the durable side of that job. That is a different job, not a next step. It is a pay cut, in those words: $74,510 against your $104,110, 28.4% less (OEWS May 2025 (both)). Retraining to earn less is a decision, not advice. And it is a narrow door: about 36,190 of those jobs against 296,810 of yours (OEWS May 2025), 12% as many seats.
Look at that job’s page anyway →
Petroleum Engineers
Why it looked obvious: It came up as a near neighbour on the work AI is not taking: you already confer with engineers or other personnel to implement operating procedures, resolve system malfunctions…, and their equivalent is to confer with scientific, engineering, and technical personnel to resolve design, research, and testing…. Across both published task lists that is about 2% of the durable work in that job.
Why I am not recommending it: Almost none of it is work you already do: about 2% of the durable side of that job. That is a different job, not a next step. And it is a narrow door: about 18,060 of those jobs against 296,810 of yours (OEWS May 2025), 6% as many seats.
Look at that job’s page anyway →
Industrial Engineers
Why it looked obvious: It came up as a near neighbour on the work AI is not taking: you already establish production or quality control standards, and their equivalent is to evaluate manufactured products according to specifications and quality standards. Across both published task lists that is about 2% of the durable work in that job.
Why I am not recommending it: Almost none of it is work you already do: about 2% of the durable side of that job. That is a different job, not a next step.
Look at that job’s page anyway →

What I’d stop worrying about

A friend tells you what not to spend fear on. This is that list.

The headline number you read somewhere
The big “X% of jobs” figures are about the whole economy, not about you. The number that describes your job is on this page: 19% of its task weight, across 79 scored tasks. Every row behind it is printed above with the source. Check ours; ignore theirs.
The headlines about your trade disappearing
They are usually about the technology, not the timetable. Changes to work like conducting or directing system-level automotive testing arrive through rules, insurance and money, slowly and visibly. This page tracks the task evidence and will move when it moves.
Retraining out of a job that is holding up
On this evidence I would not spend money leaving. Spend it on the layer around the job instead: the tools, the paperwork, the planning. That is where the change actually is.
The “obvious” next job everyone suggests
I checked the obvious moves and most of them did not survive. The reasons are printed with the routes above, including the pay and the gate. A move that fails on the numbers is worth knowing about so you can stop turning it over at night.

You are reading the United States figures

The United Kingdom splits this work across more than one official group, of which Mechanical engineers (professional) is the closest. The pay and employment figures are not directly comparable, and we do not average them together.

Switch to the United Kingdom page →partial match

The other groups this work is counted across:

Aerospace engineers

In UK official statistics this job is counted as Mechanical engineers (professional) and Aerospace engineers. Pay is shown separately for each of those groups (medians cannot be averaged together), while the task list and the scores on this page are for this group only.

Your route through this

Where to go next, and what it costs

Free, and complete

The moves above cost nothing. These are the real services that go with them: public, government-funded, and free at the point of use. Nothing on this page is behind an email address or a payment.

Anywhere in the US:
CareerOneStop - Find local training
Search what's running near you, from the Labor Department's own database, before anyone sells you a course.
Free to search; individual programs vary, and some are funded
Anywhere in the US:
American Job Center locator
An American Job Center will sit down with you for free. Find yours by ZIP code.
Free
- CareerOneStop local help hub
Anywhere in the US:
CareerOneStop - Licensed occupations finder
Check what your state actually requires before you pay for anything.
Free
Anywhere in the US:
CareerOneStop
Free

No Space for this job, but one for what is happening to it

Nothing Collab365 runs is built for mechanical engineers, and we are not going to point you at the nearest one and call it a fit.

There is one that is not about a job title at all. The AI Authority is about being the person who directs these tools at work rather than the person they get compared to. That is worth saying here, because 19% of the work on this page is already inside what they can do.

Try The AI Authority free →

7 days free, no card needed. Explore up to 2 Spaces before you choose a plan: you pick a plan later, not now.

The AI Authority is a general community about working with AI, not a course for mechanical engineers. You do not need it to act on anything here: the moves above cost nothing and stand on their own. The data on this page is the same either way.

Noted, and thank you. We’ll email you if a Space for mechanical engineers launches. Nothing else.

That did not look like an email address, so nothing was saved. Have another go below.

We could not save that. The fault is ours, not yours, and your address was not stored. Please try again later.

No Space for mechanical engineers yet. Should there be one?

Collab365 Spaces is built by the same people as this site. We find the problems that AI and automation are creating inside one kind of work, then solve them as short courses, briefings and Blueprints. Each Space is the community too, so the research and the people doing your job are in the same place.

What a Space actually is, in full

Collab365 launches new communities where the need is real. If one for mechanical engineers existed, with researched problems, courses and people in the same boat, would you want in?

Your email

We use your email address for one thing: to tell you if a Space for mechanical engineers launches. We never sell it, never use it for unrelated marketing, and every email has a one-click unsubscribe. Our privacy policy has the full version.

This unlocks nothing. Every figure, every row and every step on this page is already yours, whether you fill this in or not.

No deadline on any of this. The page will still be here, and the data is refreshed on a published schedule rather than when someone wants a headline.

Questions people ask about this job

Will AI replace Mechanical Engineers?: Not as a job, but it is already doing parts of the work. Across the 79 official task statements scored for Mechanical Engineers (United States, SOC 17-2141), 19% of the importance-weighted core work is made of tasks today's AI could already do most of. The overall exposure score is 37 out of 100 (range 31–44, band: low). That is a statement about tasks, not about headcount: this measures what AI could do, not whether any employer adopts it, whether the law allows it, or whether doing the routine parts faster creates more demand for the human parts. Figures are from release 2026-q4.1.
Which tasks in “Mechanical Engineers” can AI already do?: The highest-scoring tasks in release 2026-q4.1 are: “Write, review, or maintain engineering documentation” (81/100, very high); “Prepare or present technical or project status reports” (79/100, high); “Research computerized automotive applications, such as telemetrics, intelligent transportation systems, artificial intelligence, or automatic control” (75/100, high). Each score comes from five published 0–4 ratings turned into a number by a published formula, and each carries the model's one-sentence reason on the page.
Which tasks in “Mechanical Engineers” stay human?: About 58% of this job's task weight sits in work that scores low for AI exposure. The lowest-scoring tasks in release 2026-q4.1 are: “Prepare test stations, instrumentation, or data acquisition systems for use in specific tests of fuel cell components or systems” (0/100, minimal); “Fabricate prototypes of fuel cell components, assemblies, stacks, or systems” (0/100, minimal); “Conduct fuel cell testing projects, using fuel cell test stations, analytical instruments, or electrochemical diagnostics, such as cyclic voltammetry or impe…” (7/100, minimal). Low scores usually mean the task needs a body in a room, a legally accountable human, or trust built in real time. Those are the three things the scoring rubric treats as gates rather than obstacles.
What should someone working in “Mechanical Engineers” do about AI?: Start from the ledger rather than the headline: 19% of this job's weighted core work is exposed, and roughly 58% is not. The practical move is to spend more of your week on the tasks that score low, the ones above, and to get fluent at directing AI through the tasks that score high, because those are the parts that change whether or not you are ready for them. This page does not predict your job, and nothing here is career advice tailored to you: the score describes the occupation, not the person.
How is the AI exposure score for Mechanical Engineers calculated?: Each official task statement for the occupation is rated on five published 0–4 dimensions (output replicability, physical embodiment, licensed accountability, real-time human trust, and data availability) by claude-opus-5 using scoring prompt task_scoring_v1.0. The model never writes the score; a published formula turns the five ratings into a 0–100 number, so every score can be recomputed by hand. The occupation figure is the importance-weighted mean across 79 scored tasks. The prompt, the rubric, the formula and the full dataset are published at https://futureproof.collab365.com/method and https://futureproof.collab365.com/data/2026-q4.1 under CC BY 4.0.

Where these numbers come from

Worth knowing about these figures

The match between this job and its counterpart in the other country is partial, so the two pages count slightly different groups of people.
8 of this occupation's scored task statements carry a score that was measured under a different occupation's context, because the statement is shared between jobs and has only been scored once so far. Each one names the occupation it was measured under in the free routes below; none is presented as a measurement for this job.
Task clusters are not derived in this release, so the task-cluster field is empty and no Collab365 Space routing is attached to this occupation yet.

Task statements: onet-dbProcessing: catalogue-bridge → onet-im-rt-weighting → task-scoring → score-aggregation
Task weights: onet-db (im-rt)
Scores: Rubric task_scoring_v1.0, model claude-opus-5, computed 2026-08-04.
Pay and employment: bls-oews (May 2025 estimates (national_M2025_dl.xlsx))bls-oews (May 2025 estimates (national_M2025_dl.xlsx))

Figures on this page come from release 2026-q4.1, published 2026-08-05. Every release keeps its own permanent address, so a figure you cite in March is still there, unchanged, in November.

The plain-English wording on this page is assembled directly from the task statements and the published ratings, not written by hand for this occupation. That is why it is specific, and it is also why we say so.

The routes and free resources further up are today’s, not the release’s (last reviewed 2026-08-05). A route is an offer, not a historical fact, so it moves on its own clock.

How we score a job Download this release Look up another job

Using these figures?

Cite this

Everything on this site is published under CC BY 4.0. Quote it, chart it, sell something built on it. Just say where it came from, and cite the dated release rather than the site, so the figure you quote stays checkable.

Plain text

Collab365 (2026). Collab365 Futureproof: task-level AI exposure for US and UK occupations, release 2026-q4.1 (methodVersion 2.0.0, promptVersion task_scoring_v1.0). https://futureproof.collab365.com/data/2026-q4.1. Licensed CC BY 4.0. Built with O*NET data (USDOL/ETA, CC BY 4.0); ONS data (Open Government Licence v3.0); GAISI task framework (arXiv:2507.22748, MIT); BLS data (public domain).

BibTeX

@misc{collab365futureproof2026q41,
  title        = {Collab365 Futureproof: task-level AI exposure for US and UK occupations, release 2026-q4.1},
  author       = {{Collab365}},
  year         = {2026},
  url          = {https://futureproof.collab365.com/data/2026-q4.1},
  note         = {Release 2026-q4.1, methodVersion 2.0.0, promptVersion task_scoring_v1.0, CC BY 4.0}
}

Permalink

https://futureproof.collab365.com/data/2026-q4.1

Data as of release 2026-q4.1, published 2026-08-05. Releases never change after publication; when the figures move, a new dated release is published beside this one and this one stays exactly where it is.