I came across an ad for Home Depot’s “Path to Pro” microsite, which seeks to recruit “future Pros,” and an interesting question came to mind.
Is Generative AI Going to Drive More Workers to the Trades?
Around 20 years ago or so, one of my uncles gave me a piece of advice about career paths. I remember wanting to steer towards aerospace engineering; if I couldn’t be an astronaut, maybe working on spacecraft would scratch that itch. He asked me – are rocket engineers in high demand right now? No, they weren’t.
Advertisement
I learned broad skills that I continue to put to use on a daily basis.
The mass media keeps pushing the idea that generative AI is going to decimate all kinds of white collar industries. It might.
Generative AI can put together marketing copy, press releases, news articles, and even reviews, albeit truly terrible ones from what I’ve seen.
AI doesn’t have talent, creativity, or originality; generative AI cannot form unique ideas. It cannot think or problem-solve.
Generative AI takes an input and applies patterns to it. It’s… algorithmic.
But still, there could be some truth in the many predictions about AI’s impact on a lot of types of jobs.
Advertisement
I think that most trades jobs will be safe. Whether white collar or blue, skills and experience will always be in demand.
A lot of trades professionals make decisions, and also draw upon experience to troubleshoot and make judgements. For now and the foreseeable future, this is outside the realm of generative AI.
Stemming from that idea, I think that job security – if trades jobs are indeed safe from generative AI and next-gen robotics – is going to be a strong recruitment tool.
What do you think generative AI will mean for the trades?
Thoughts?
teicher
I think it could help with administrative tasks, such as generating drafts for emails, quotes, tc.
It could also be embedded into some of the scanning and digital tools, to help analyze and diagnose systems
grannygearga
I think that the trades will be most affected as a result of upstream use of AI. I am no expert on the design tools that residential architects use, but in my limited experience modeling a house in software is still a very manual,slow, and incomplete process. The builders and trades are expected to bring their experience to bear on the details of the implementation.
But, if architects leverage AI to provide builders with a *completely* modeled home, for example, that changes things dramatically.
For example, the model could include precise details of every single circuit that needs to be run, all the way down to which studs need holes drilled in them to optimize the wire routing. It wouldn’t take much training or experience to follow those instructions.
Doresoom
Ha, that sure would have been useful when drawing up my basement plans last year. I got lazy and put a note “electrical outlet and lighting placement to be determined by licensed electrician.”
The city permit office rejected it so I had to place outlet, can light, and CO/smoke detector locations on the plans to get them approved. It would have been really nice if there was software to just place outlet diagrams every 12 ft automatically.
After a discussion with my electrician, we agreed that he should just place them where he thought was best anyway. 🙄
Rafe
Typical residential architect “mill” firms use a lot of pre designed house plans as it is and are equally in demand for creating site plans as they are house plans. The site plan is a little bit more complicated because it’s subjective to different code requirements and differing qualities of surveys to create the site plan.
Having a house plan that is too detailed for a build that isn’t particularly complicated would be mostly useless. A lot of guys barely look at the plans as it is. Also, Auto CAD is pretty fast. You’re still going to have engineering requirements on top of your house plans anyhow and I suspect AI won’t have an engineering cert stamp any time soon.
These jobs are pretty safe I think.
S
That works great in concept for spec homes. Any custom homes or remodels, you might as well throw the whole computer out.
Jared
Are you SURE about that? That sort of sounds like something AI could do to me.
I’ve experienced AI trained on a complex set of rules and then asked to answer very specific questions about what is allowed where. It was remarkably good at it – and way faster than a human at coming up with the answer.
You might notice in my earlier comments farther down that I’m not all that enthusiastic about AI – but AI can synthesize information, not just apply a set of rules. This application might actually be playing to its strengths.
Samuel VDot
I can see how Generative AI would be useful for analyzing service calls. Is there a consistent part that is failing? Is model Y failing consistently after Z years? If so, generate a sales touch point for an upgrade before Z years. Is Employee X ‘s work causing increased service calls?
It could also be helpful in analyzing changes in building codes or other regulations.
Jared
It’s a good marketing line in my opinion. Whether it actually bears out? Can’t tell yet.
A lot depends on how much it continues to improve. I’ve heard much speculation about how it will disrupt my white collar job, but even though I’ve made some effort to incorporate it into my work (don’t want to be left behind after all) – it’s just not that great.
For example, the current generation of AI is a tremendously mediocre writer. No matter how I prompt it, it seems to embellish its content with at least 25% more words than necessary.
It’s also been heralded as a fantastic way to condense material and pull out the key points. Sometimes it can, sometimes it misses the point or “hallucinates” and makes stuff up. You need to check its work.
The one thing I’ve found very useful is as a brainstorming tool. Ask it to answer a question and it can make a list of things that might be relevant considerations – the technically skilled human still needs to look at the answer and decide what is actually important. Nevertheless, even when you know a subject well it can take some effort and time to make sure you haven’t omitted anything.
All that to say, I’m not sure yet just how far the “white collar” disruption will go. On the one hand, this is the worst version of AI we will ever see – it’s going to improve. On the other, it’s fundamental nature seems to impose limits.
I think it’s bound to be a productivity booster, as it improves and people learn how to use it, but does that actually mean slews of people will become unemployed? The internet helped speed up research, did that mean that many white collar workers became superfluous because they were suddenly more productive?
Tom
At least in law firms, Word processing caused mass redundancies over the course of a few years in the 1990’s.
My wife’s first law job (1994?) saw her with a PA and a company typing pool. Both gone from the trade by 1997.
MFC
My brother works for Amazon as a software engineer. He’s been doing this for over a decade. He, and all the software engineers he knows, go to google and forums to find fixes and patches for their software all of the time. Generative AI will be very helpful to them to increase productivity, which would mean that they don’t need as many engineers to accomplish the same tasks, but this is true of all industries. As technology increases, productivity can be increased as well. I don’t see AI taking their jobs yet, but it is certainly helpful. So, I think AI will be extremely helpful for some who use it to create more content, or tailor things more quickly, but it’s not human, as you pointed out, and can’t run automatically. Yet.
Ten years from now? Yeah, it’ll do a lot more, but it will open the door for more jobs as AI will be creating new markets that never existed just like the internet did.
Michael F
I’m a software engineer and I absolutely believe generative AI will reduce demand for coding skill sets. In fact, the startup I’m working at now provides a service that fixed security vulnerabilities in code using gen-AI. The code that gen-AI writes is actually pretty good.
MFC
Yeah, neither he nor I have any doubt that it won’t fill a lot of holes in the coding department, but you prove the point because you literally have a job working for a company that is utilizing AI. Some people in your company might not have jobs without it. Also, fixing code and designing code are different. Eventually it will be able to “design” based on parameters that guys like you give it, but we will still need guys like you.
Hon Cho
Many of the things listed as shortcomings of AI are inherent flaws in a great segment of the population. Many of us like to think we’re not susceptible to manipulation but the vast majority of us are susceptible to one thing of another. AI is likely to increase the manipulation of us all and increase the speed to the bottom in so many segments of society. Pre-digested “information” and entertainment served up just the way that tickle our brains in a way to empty our wallets, sway our opinions and votes.
Sure, there are benefits to AI but it’s just not clear to me that they will worth the costs that society will bear.
As for for trades, maybe AI will make people realize that if they want a decent job, they’ll have to do something that is valued by others and can’t be replaced by a machine. Probably be a long while before plumbers and HVAC techs will be replaced by AI powered robots.
Steve
Can’t see AI affecting plumber or HVAC for a very long time. Not in my lifetime anyway…
Mark M.
AI is (and will be) disruptive, but it’s starting to smell like virtual reality and the Metaverse. Lots of hype but largely underwhelming at the moment.
Wayne R.
C’mon – it’s the 3D TV of the future!
Jason
Just wear your glasses and try not to get nauseated.
Steven+B
Generative AI is BS!
I’ve worked with “AI” for a decade now. Other facets, like ML and Machine Vision are legit, but the hype around Generative AI and Chat GPT is NONSENSE. It is nothing more than fancy autocomplete with a high carbon footprint.
If you’re paying a human being to do it now, it either has to be correct (like a legal contract) or as compelling as possible (movie script or advertising copy). Chat GPT has no clue what it’s doing and is incredibly inaccurate and scientifically has no realistic chance of ever being accurate enough to match the intelligence of the dumbest human beings. It has no clue if it’s writing a product manual or a story about mermaids and regularly mixes content.
ChatGPT is fun and whimsical, but not accurate. It’s main use case now is for entertainment, like when I had it write my 8yo a bedtime story about her best friend being a mermaid with her favorite Disney Channel characters…or for scams…generating content that looks correct enough to fool someone for a phishing site or investment scam.
If you want to generate content today and you don’t care if it’s accurate or compelling, like my daughter’s bedtime story, you aren’t paying money to have it done today.
If you’re paying money to have something written, you’re doing so because a mistake will risk a lawsuit or decline in sales.
The collapse of globalization, China’s implosion, and the migration of manufacturing from a global supply chain to a regional supply chain is likely to increase demand for the trades: both because a lot of factories are being built in North America and also because a lot of working-class jobs in these factories are opening up…drawing some out of the trades.
I am skeptical Generative AI will cause a tangible net loss of jobs in the next 20 years.
eddie sky
Trades don’t depend on Home Depot, they depend on the paint store, the lumberyard, the window/doors specialist, the plumbing supplier, the electrical suppliers.
When you are in the trades, and say a contractor/general contractor, HD and Lowes are your failsafe or last resort. Their lumber and materials vary and QC is poor. Even their delivery can be late (a no no) or not quality picked. And alot depends on the store management for that region.
I’m not worried too much on AI because what you need to know is, good writing and math for it to work for you. Generative is still a system that scrapes or takes in and then spew out, on specifics. You feed it shit, it will show that. And with copyright infringement suits just starting, I think for some, it might help save costs, but also kill those we don’t need like sales, marketing and copy work (which we know is just entrapment prose to sell/market you something).
Jim Felt
Of course these folk’s (endangered) careers presently lead to a lot of middle class incomes wanting/needing the trades to update and/or expand their homes.
Always the absolute Law of Unintended Consequences.
James
This.
Jim Felt
Stuart. Perhaps we’re/you’re forgetting “Colossus: The Forbin Project”? Or more recently “Skynet”?
The future is called that because it’s unknowable.
I clearly remember people saying cell phones were silly in the mid ‘80’s. Apple would fail at the turn of the century and we’d never have a common retail pocket devise more powerful then the mainframes that sent the astronauts into orbit.
BigTimeTommy
It’s going to be used to drive wages down and screw labor, of course. Unskilled, underqualified people will be sent out into the field with AI assistance to do all the thinking for them. Of course the AI assistant will be incredibly flawed, but we live in a globally collapsing society where almost no one cares anymore. Sorry if this is “politics” stu.
JoeM
Generative AI, in regards to Tradecraft, can only do one thing. Generate things for Tradespeople, and their associated companies, to build. It may (note the modifier: may) create more challenging builds than we’re used to seeing, or it might just very easily design ready-to-build kits for efficient work among those in the Trades.
If an architect can’t get a window to do what they want it to in CAD, perhaps they turn to their company’s Architecture-Oriented AI to get the look right? “A Glass Window That Refracts Fractals On The Walls When The Sun Shines Through It, But Is Entirely Transparent and UV Blocking Otherwise.” Is anyone expecting an Engineer or Architect to manage to embed the Mandelbrot Set into glass by hand? No… I doubt that very much. Do I expect that product to be hung in its framing by a Tradesperson? Absolutely.
All this fearmongering about AI is a total overreaction. AI can’t replace Creatives. “New” ideas need to come from humans… Maybe Dolphins or the Great Ape family, if you want to get funky. The worst that will come for the Tradespeople is everyone staring at the blueprints and build plans, and thinking the Designers and Architects need heavy medication and therapy. Really, that’s it. Building as usual. At worst, operating, programming, and supervising, a machine that does the job they used to do. It’s less likely to be like Ford converting the Automobile or Coach building process to an Assembly Line with half the employees, and more likely to be like someone taking a screwdriver out of someone’s hand, and putting a drill with a bit in its place. Same workers, same skills, new tools.
AI is only a hazard in movies.
MFC
It’s not the AI that is a hazard. It’s what people do with the capabilities. People are right to be afraid of it because it gives the “ruling class” the ability to control. AI is a puppet, but it’s a puppet that will continue to develop new swiss army knife functions.
As far as it taking jobs? I agree with you. “Same workers, same skills, new tools.”
xu lu
You are correct re creatives. However true creative is about 5 pct. The rest is non really creative but derivative. AI will gut 95 pct of ‘creative’ just like nearly everything else it will kill. The systems we have designed to improve productivity have created a copy cat culture whether in pharma or media etc rendering the coming disembowlment of employment a walk in the park for AI.
Derek H
I’m well-versed in this topic, so here’s my stance.
Gen AI will impact the white collar side of the trades, waaaay more then the people who do the actual work.
AI is getting good, at a staggering pace.
People who work in the field will get replaced by embodied AI (aka: humanoid robots). This isn’t 20 to 40 years off, but far sooner.
The number of companies working on humanoid robots is constantly growing, with the leader being Tesla, with Optimus Gen 2.
UBI will be passed by congress in 2030, with many people retiring once they start getting their free money deposited into their account every month.
They won’t have a choice but to pass UBI, because by then AI and humanoid robots with AI, will be replacing people by the millions. If a company can buy a humanoid robot for 60k, and they pay you 60k, the robot is by far the smarter decision.
I’m a 15+ yr electrician, and I think I’ll have a robot apprentice in ~5 years, and be retired in 10 years. Not because I have millions saved up, but because I will not be able to compete against a robot that is smarter then me, works faster then me, never takes bathroom breaks, or gets sick, doesn’t have health insurance, retirement account, it will always be ready to work, & never complain. It’s only downtime will be very minor maintenance, and recharging 1-2 times per day. Likely meaning it works 20 hours a day on average.
These are general purpose robots, not some big robot arm that can only do one task, these will do nearly everything.
Oh, and please don’t comment on saying Boston Dynamics doesn’t seem THAT impressive. Cause if you think they are the best out there, then you must also think Adrian Peterson is currently the best running back. You’re living in the past bro.
Jeff
Boston Dynamics was working with DARPA. Those robots were what they felt comfortable sharing with us. What they have behind closed doors in some black budget program, is likely 30 years more advanced.
Yes, the A.I. & robots are advancing… but we only get the trickle down… will we ever get to see the good stuff they been hoarding for decades? (Insert “raiders of the lost ark” closing scene)
Derek H
No sir. BD has no secret tech behind the scenes, they are a private company, currently owned by Hyundai.
Atlas (their humanoid robot) is pre-programmed, and hydraulically controlled, very inefficient, and expensive.
Spot (their “dog”) is crap, has one arm with claws, and struggles to turn a door knob.
Stretch (their semi unloader) is only capable of doing that one task, and has no future doing anything else.
A few months ago I watched an interview with their CEO, and he talked about Atlas, and how it’s just for R & D, and they have no plans to ever commercialize it.
Only the uninformed think they are the best.
They still have some talented people working there, which is why I rank them #9, on my current list of 12 companies working on humanoid robotics.
Stuart
Self checkout was supposed to replace labor and increase shopper convenience. How well did that work out?
KMR
Self Checkout is fantastic and I’d say it benefits both the business and the willing consumer.
1) Our Wegmans grocery store has 12 self checkout spots staffed by 1-2 store helpers/observers depending on how busy the customer traffic is. So there clearly is a labor reduction for the number of customers able to be checked out.
2) With 12 self checkout stations I rarely have to wait to checkout, with only busy holiday periods or seasonal times, having a small wait to get to one of the spots.
3) The actual checkout process is quicker for me to use self checkout than a standard cashier, unless I have bulky items (paper towel, etc) or any item that requires the intervention of the helper/observer.
4) I don’t have to interact with anyone and no one but me has to touch the stuff I’m buying.
Tom
I’d concur with you, self checkout has worked out really well here in New Zealand. Practically universal in supermarkets, growing in large chain stores.
NoDak Farming
I’ve play around with ChatGPT, and read articles about it. And I think a lot of the above comments hit the nail right on the head. I think the very first comment about AI helping with administrative tasks is a perfect example. More and more people are working for themselves. And even now already, free versions of ChatGPT are helping put together better reports, more professional looking estimates, neater bills, and tidying up business finances. This will help those that didn’t go to business school, but are darn good craftsman. It takes some practice and trial & error experience. But people are going to be able to leverage text based AI for their personal business without having to take night courses.
I also agree that it’s a type of new tool. At one time, if I needed to research something, there was the library and the stack of out of date encyclopedias in the living room. Then came the internet. But you still had to know what website would even have the right info. Then of course Search Engines streamlined all of that. And now we have text based AI that streamlines things even further. It might turn into something crazy. But I think it can be a useful tool if used right.
Brad
I think the intersection of robotics and AI will eventually displace some tradespeople, starting on large scales (e.g., office building construction as opposed to bespoke home renovations), but probably not in the short term.
On the other hand, I’m pretty sure it could be a huge enabler of small business, making it much easier for your average tradesperson, artist, or consultant to strike out on their own. I can’t imagine it will be too long before we’ll see off-the-shelf subscription software/service suites that enable all types of workers whose talents lend themselves to small enterprise, but who lack the patience, knowhow, or administrative conscientiousness to actually operate independently. More time making, doing, and selling, and less time futzing with business and tax records.
This could just be wishful thinking, but I can’t be alone in wanting to make, market, and hawk my widgets, without the stress and administrative burdens of book keeping, record keeping, quarterly taxes, etc.
Scott K
Totally. I know I’m not alone in having had a good experience with someone’s work but a bad experience with the admin side of their business. I can think of multiple occasions where it took so long to get a quote that I moved on or received an invoice or had a check deposited so long after a completed job that I questioned how they could sustain that timeline. Finding the time for this type of task is a challenge that seems apt for AI or other software.
xu lu
In say 25 or 30 years the trades just like nearly every other occupation will lose 50 pct of workers. Complete and total annihilation of jobs large and small is coming rapidly. 20 to 25 pct non govt white collar gone by decades end absent political intervention.
AI coupled with tireless robots will diagnose and perform the work. Aided by AI, products will be redesigned to play to AI/robots ‘strengths’. The next target of the green revolution will be homes perhaps including individual ownership and this will drive construction of AI/robot friendly environments. While in the 60s Motorola had a failed color TV line with the works in a drawer to facilitate repairs, this is the future of many systems. Because of the existing infrastructure investments, the trades have a better chance of holding on longer, than others but substaintially diminished they will be. Robot roofer, iron worker, lawn sprinkler installer etc etc. Yea it’s the tobots plus AI that should be feared.
Derek H
Fearing robots makes no sense, that’s progress, it’s inevitable.
The fear comes from money, no job = no money. no money = no…anything.
But UBI changes that, our lives will get much better if we embrace this change, and it will be a life of leisure.
50% of all jobs will be gone by 2030.
Jimi
50% of “all” jobs gone in 6 years , no .
Doug
2030 is only six years away. There’s nothing coming that fast that will eliminate half of all jobs.
You seem to be caught up in the hype. AI research in its current form has been happening since the 70’s. Even the neural net model that’s gaining all the attention in the current news cycle is over a decade old.
Derek H
I’ve been following Robotics for decades, and AI didn’t truly show up until November 2022 with ChatGPT.
The hype is for a reason, a tsunami is building, and few seem to realize the water is receding.
People who pay attention know that the work on AI and humanoid robots are finally converging. Having embodied AI smart as a person, changes everything.
The growth will be HUGE.
Global humanoid robot numbers:
2023 – ~300
2024 – ~5,000
2025 – ~ 150,000
2026 – ~ 500,000
2027 – ~ 3,000,000
2028 – ~20,000,000
2029 – ~ 100,000,000
2030 – ~300,000,000
In 20 years we will have more humanoid robots, then people on Earth.
Lets keep in mind this is just humanoid robots, so it doesn’t count self driving vehicles, and the biggest job in the US…is people driving.
Also doesn’t account for all the office jobs where AI doesn’t need a physical body. Accounting, Secretary, Lawyer, Tax preparer, Telemarketer, Customer Service Reps, and many other data entry jobs.
Stuart
Generative AI isn’t actually “AI” . It’s feeding an algorithm countless copywritten works until it can spit out different combinations of words in ways that resemble human language patterns.
Predicting growth from 300 to 300 million “humanoid robots” in 7 years is unrealistic.
S
I think I saw that movie.
You have a lot more faith in robots solving problems than I do.
Does anyone else remember all the propaganda in the 1970’s where they kept saying by 2000, we’ll all have flying cars, robotic dinner makers, and never need to work another day? Where’s that future at?
Reality is always a bit more bleak than the ad’s make it out to be.
I don’t believe self driving cars can happen in the next 20 years simply because of the legal ownership and status symbol issues.
It’s more profitable for the OEM’s to stop at level 4 systems, where the current liabilities and status symbol ownerships fall onto a party other than the OEM.
Humans are unique problem solvers, and there’s a reason that robots aren’t currently used outside assembly lines. The world and every trade has tons of problems that need solving on a daily basis that robots and strict rules simply can’t allow for
Derek H
“Generative AI isn’t actually “AI”
It’s in the name for a reason, it absolutely is AI.
It’s machine learning that can create new things that don’t exist, whether that’s a picture or video, or code for a app that you thought of, but you don’t know how to code. It will do it for you.
That tech is being put into humanoid robots NOW, hell, just yesterday Figure showed off a video of their robot making a cup of coffee, from a simple voice command.
I’m not trying to de-rail your tool site, I love tools, I’m just trying to stress that people should know what’s on the horizon.
Doug
News flash. Chat GPG has been around since 2018.
Derek H
Doug – ChatGPT 3.5 was November 2022, that is when the history books will say AI “arrived”. Previous LLM’s weren’t large enough to be good. 3.5 is when it crossed over to being a valuable tool.
Brad
It’s a little bleak, isn’t it? If we don’t find ways to adapt our economic paradigms to worker replacement, eventually most of our kids are going to feel like Youngstown’s steel workers.
Scott K
I disagree with some of this. AI and robots do not handle “chaos” well. Sure, there will be a time when some simpler tasks may be handled by robots like painting a prepped surface. But there are several steps involved in roofing and most houses are not built perfectly which often requires adapting to newly uncovered issues or other unknowns. My school district purchased robotic field liners and my understanding is that they require a “handler” to set them up, manage them, and assist when they get stuck. They came with a very hefty price tag.
Scott K
Not to discount the potential impacts of AI, but I think other factors will influence the trades in the shorter term. I think that the pendulum will swing and trades will become more attractive as the shortage of skilled workers becomes even more significant coupled with the cost of college reaching an even higher level. We may even see schools adopt trade programs or partner with trade schools. Several schools offer co-op programs and I think we will likely see more creative uses of this model.
Mopar4wd
I think the simplest thing is to look at history. As we have automated more and more things do we lose some jobs , well yes, but they have always been replaced be something new we find for people to do. Economist have predicted large falls in working hours from automation in the past and it just doesn’t happen long term. We just find something new to work on.
Nathan
Wow this devolved into a ubi discussion. Where will that mythical money come from when everyone is on poverty government allotted pay?
Meanwhile I’m tired of hearing ai. Anyone remember when it was called fizzy logic? It was the first version.
One thing intelligent programing will do is make the mundane parts easier and maybe more efficient. You see it today in metal fab with the laser cutters that optimize the geometry of the cuts to minimize sheet loss. It’s not new either
Robots that can do actual work are coming I suppose there are some now with assembly. It has to be a fairly controlled process and repeatable. I mean I don’t recall the last time a car body was welded by hand but it was probably in the mid 90s
Derek H
” Where will that mythical money come from when everyone is on poverty government allotted pay?”
It’s probably gonna be a messy transition, but at first, that government allotted money will be coming from taxes…same place it does today. They will be taxing companies based on how many robots they employ. It’s the only logical way to do it. Everyone over 18 will get a check, IDK how much, but my guess would be at least 2k per month. Which for some would be a raise, but for many it wouldn’t be. again, messy transition. We have never moved into a jobless economy before, and it will not go smoothly.
Stuart
This isn’t the place to talk about the potential for UBI.
You’re getting too far off-topic and maybe crossing over the no-politics line at the same time.
Nathan
One of my favorite features of fusion360 is having it make 2d prints of my parts of a 3d design. Handy for the home builder
Cullen J Webb
As somebody who uses AI every day to augment my projects I want to start by saying that it’s not as good as people may think. Human creativity is still the glue that pulls all of the different moving parts in any project together.
However, I think that you’re on to something. If enough people only *perceive* AI as taking over certain industries that could very well influence any decision making about their future. Even if AI never improves beyond what we see today I wouldn’t be surprised if in 5 – 10 years we see a major shift towards the trades. A blessing in disguise?