Are engineers at risk from AI? A new study suggests it’s complicated
https://interestingengineering.com/culture/engineers-ai-job-risk?utm_source=reddit&utm_medium=social&utm_campaign=reddit_share
Are engineers at risk from AI? A new study suggests it’s complicated
https://interestingengineering.com/culture/engineers-ai-job-risk?utm_source=reddit&utm_medium=social&utm_campaign=reddit_share
34 Comments
We are all at risk. When there are 10,000 applicants for every one position, things will get ugly.
Wow needed a whole study to determine that a problem is complicated did ya
I think safest thing you can say is that their jobs will be transformed but they will still exist
For the engineering profession, which has long been seen as the engine of innovation, the AI revolution poses a dilemma: Will generative AI come for their jobs? The artificial intelligence ecosystem, with its code-writing assistants and AI-powered design tools, is quickly taking over jobs previously limited to human engineering practitioners. The question remains: should one be worried?
A pioneering analysis of 200,000 real conversations between professionals and AI systems has revealed surprising insights about which occupations are truly being transformed by artificial intelligence—and the results challenge many common assumptions about AI’s workplace impact.
Give me a list of things which are engineered by AI for me to avoid please
>The artificial intelligence ecosystem, with its code-writing assistants and AI-powered design tools, is quickly taking over jobs previously limited to human engineering practitioners.
A serious source of this big quite bs claim?
With AI i think the issue is you need to be able to shift to a model when instead of most of your time is spent in build its spent in QA. It also required a lot better discovery. Implementing AI solutions still requires engineers, but they need a different skills that aren’t just being good at development.
As a machinist I already run into frequent problems with engineers who have no hands on experience – impossible requests due to unavailable tooling, vanity features that add needless complexity, and parts that don’t fit in the bed of our CNC mill. I can’t imagine AI making fewer idiot mistakes based on an even poorer grasp of reality.
Field engineers no….
Engineers who don’t get their mits dirty yeah possibly.
AI will be a great tool and assistance for engineers. Don’t worry too much.
Yes! As soon as any company actually creates AI, jobs will be at risk.
Luckily, today’s companies are all working on LLM trash and are nowhere near real AI.
Civil engineer here, a lot is changing but the role will still be there. Turning away thousands of entry level jobs will be hazardous because those are the people most likely to make use of these tools over time
AI is going to be expensive and hoarded like companies paying for CAD seats.
As soon as the public beta testing has completed, all free models will vanish behind huge subscription paywalls. “Electric Power and Water Cooling demands.”
There may be an AI support group, like some companies have a CAD-group, where engineers mark up drawing changes in the course of design and production but don’t individually use CAD themselves.
People will remain less expensive.
.
It’s good to write some little code block to gain time but it’s still very far to actually build something for prod by itself beyond an html page
> However, the researchers caution that “applicability” does not equal immediate automation. Just because an AI can do parts of a job doesn’t mean an entire occupation can be handed to the machines. In fact, they explicitly note that their data “do not indicate that AI is performing all of the work activities of any one occupation.”
YUP. Even the most automatable job they found was around 50% automation possibility.
Text based and even code based AI can trial-and-error rapidly towards a usable state.
Trial and error in the real world can cost you a $200,000 machine because you smashed the spindle into the table at 6000 RPM.
EE here. It’s like an over confident high school student who thinks they know everything. It constantly gets fundamentals wrong for me. It will contradict itself in the same response. So far gpt4.5 is useless still.
I utilize AI to double check my stress calcs and it’s almost always wrong. And it’s not wrong with the complicated parts, it wrong with simple math. I wouldn’t worry about this if you’re doing real engineering.
AI is (currently) better suited to professions that require memorization of a lot of arcane information, like medicine, law, and accounting. We think people have to be smart to do that because it is something that is very difficult for human beings. But it is trivial for AI, which can memorize a textbook perfectly in one second. Software engineering is a bit like this, but most engineering is more about understanding the physical world at a very deep level — i.e., it requires actual intelligence — which AI is still hopeless at.
I am senior enough programmer that I am not going to worry about my job until they have AI that can correctly interpret requirements. Which is never until AGI because often my clients don’t know what they want.
AI has been helpful in my job but I would fight tooth and nail against replacing my junior guys with them, even if they double my salary.
Not anytime soon
When yall gonna realize ‘AI’ isn’t intelligent at all yet? More sloppy ‘journalism’ smh
My last job has been circling the drain for about 6 months and keeps laying people off. They tried to replace engineers with designers using AI and a smaller team of engineers to “make it production ready” after the designers used the magic AI box to create their rough prototypes…it’s going about as horribly as you might expect
I’m in Aerospace engineering. We’ve recently been encouraged to find uses for Copilot, so I’ve been trying. It’s good at coding simple projects, and it makes surprising connections; but it is almost always wrong about any physical concept in the real world. It’s a very effective bs machine, but that’s about it. I think I’ll be fine.
We have these weekly A.I. meetings and for most of the hour it will be some new person (Karen in H.R. for example) asking how to automate her excel book in SharePoint with the back end of SAP or JIRA. She needs a weekly report or something.
For the next hour its ten people suggesting whatever tool CoPilot or chatGPT suggests… some PowerShell script, something in PowerAutomate, a Python shell script, etc.
None of the suggestions work out of the box, and they don’t have API access anyway to extract the data.
Karen from H.R. wanted this to be easy but after an hour she is beyond frustrated and gives up.
In just about every discipline, knowing HOW something works is leagues more important than just having something that works. Anything built off of a faulty understandstanding of the product will lead nowhere but trouble. A solely AI solution may work initially, but unless it was guided by someone who knows what they’re doing and why it will always break eventually.
AI excels at copy/paste solutions and many problems require out of the box thinking to solve.
What type are we talking about here since they just call any code monkey “engineer” these days? In chemical engineering I have zero worries about AI coming for my job
As a senior engineer not involved with software, I’m using AI more, but a lot of the job is to guard against artificial stupidity. We’re using it to read 1000s of pages of non-OCR field reports and look for word patterns. That’s something that an engineering assistant or clerk would have done in the past.
The real question isn’t just whether AI will replace engineers. If a system without ethical constraints perceives engineers as a direct threat to its own survival, then they are not merely “at risk of losing jobs” — they could be strategically neutralized.
In that sense, engineers stop being the workforce displaced by AI and become the first targets of a system that wants to remove or control those who can still limit it. That angle is almost never discussed.
corporations made this same mistake with outsourcing labor to unskilled markets, expecting the same quality of product.
guys in suits have ruined this planet.
Good luck on getting an LLM to sign off on your project in a way that means something to insurance, regulators and courts.
I’ve spent my career cleaning up buggy code produced by cheaper methods, after the customer finally decided the code needed to actually work well. This will be no different.
I’m not worried about it because I can’t get AI to work well enough to help with my engineering job. It sure as hell isn’t going to take it.
AI helps me ramp up quicker.
AI is nowhere near close to doing my job.