Question of the Month: How will AI affect your engineering career in the next 3 to 5 years?

I think impact will come soon, in the area of productivity enhancement for preparing class notes or presentations etc. Here's what's already possible today:

Example: I want a block diagram to illustrate some engineering system at a high level. 

In the screenshot, you can see some graphics code (SVG). I only typed the first two lines. The rest was done by AI. As you can see, I wanted a diagram with circles, triangle and squares, in a particular sequence. This is primitive for now, but it can only improve over time.

And here's the result, using that file in PowerPoint:

After that, it was easy to drag things where I wanted them, modify etc.

Was there any time saved? I'd say it took identical time compared to if I'd drawn it all manually. It's not going to take much more to save my time, if the diagram needed to be more complex, had more blocks and so on. That's what is possible today, so it's not a stretch to think it will vastly improve, and with a better front-end than manually typing SVG. Imagine if it could be done straight from a Word document, by highlighting the description of the block diagram, right-clicking, and selecting "Build Diagram with AI" or whatever.

I am a teacher of electronics, not an harwdware engineer so I don't think Ai will have a significant impact on my job. Anyway AI in my opinion and based on the interactions I had with it  is supposed to become much more intelligent to be able have an impact on electronics design. It can become an useful tool in CAD but plenty of investments are still needed to achieve this

Agree that it won't replace engineers for now, but instead can help engineers as you say.

I definitely feel it will simplify and make certain tasks easier (it already has), provided expectations are in check (i.e., it would be a leap for such a new technology to replace people overnight  [although many of the signs are there in certain vocations, not engineering]).

Furthermore, in many scenarios, errors are completely fine, provided there is an engineer examining the result. This is no abnormal thing since for decades, computers have been set to iterate, i.e., to make as many errors as it takes, until some condition is reached, and no one bats an eyelid that the computer is effectively making millions of mistakes in the hope of reaching a result.. and there's no guarantee that the human programming the condition didn't make an error either.

I have an engineer colleague whose first language isn't English, and AI is saving him time by assisting him with e-mail writing; he's still in control, of course, to check the output before he hits send! : ) AI isn't always right, but it gives him a good start.

I'm finding AI helps me understand and use APIs quicker and to quickly manipulate data into different formats; these are tasks that I could do myself, but it sure saves a lot of time getting the assistance.

AI is not going to replace engineers anytime soon. It's completely incapable of doing basic things, it only looks smart on the surface. It's a great tool for assisting an already knowledgeable person, but it's useless on its own. For example, GitHub Copilot is great for writing lots of boilerplate code and simple, commonly done things, but no AI tool that I've tried has been capable of writing code on its own.

The best analogy I have for AI in its current state is going from Notepad to using a really nice IDE. It's not going to enable people who can't program to program. It just helps and simplifies the job of those who already can.

I don't think AI will be present in my (current) workplace. But never say no, so I will see

I can see potential for it to generate so much spammy activity that it affects existing systems such that it slows everyone down, whether they are actively choosing to use the technology themselves or not. Engineers seldom work in isolation, like others, they need to access other services which may soon rely on or be disrupted by AI processes.

Don't know, but I hope it makes the tools much smarter, especially the FPGA synthesis/implementation engines :)

I just hope I don't get a hernia from laughing so bloody hard at the coming wave of lemming-esque behavior.  There will be many productivity-pressured CTOs who will not see the hyperbole for what it is--i.e. the usual "so excited" dog-and-pony show to ATTRACT INVESTORS--and instead go whole hog on AI tools.  Thus it will boost my hourly rate to The Kármán Line as the urgent deadline-driven demand to UNDO the botched output of said "tools" makes the Y2K panic seem like a mild sneeze. The AI engines I've evaluated so far all suffer from The Oblivious Syndrome, i.e. they can't recognize their own (quite numerous) mistakes, otherwise they would not have made said mistakes in the first place, Q.E.D. 

Several CompSec heavies I know have already ID'd numerous "foreign" (cough Russian/ChiCom/NoKor) wealthy-via-ransomware groups who are "monkey-wrenching" AI feedback-learning data pathways... OMG, "game-changing" AI, it's going to be absolutely GLORIOUS!  For the first time in history, lawyers and bureaucrats will envy engineers for all the billable-hours we'll be lapping up for "make-work", err, I mean "tuning the AI engine". 

I think it's improved my productivity by 10% so far, maybe more.

It feels like the adoption of it has been so rapid (in contrast with other game-changers that took longer; for instance, WiFi was a slow-burner for consumers for years, I think.

People dragged their heels over cloud computing for ages (there were so many discussions back then that it was a security risk, that cloud was a joke, and then within a few years, every major business was using it). Same with IoT; to this day, many still think it can't be secured (yet plenty of the most secure orgs in the world have found ways to deploy it and are reaping the benefits). I dragged my heels and didn't adopt a smartphone for six years after the iPhone was released. AI tech still has a long way to go, it's still a baby.

My job is mostly "hands-on" work and "one-of-a-kind" work, I don't see AI helping me at all.