This week, I was designing a safety gate for a piece of equipment which can surely kill someone. I’m not well educated on guarding standards and we currently don’t have a person internal to the company who is an expert.
I plugged the information into ChatGPT and asked it to provide the standards for height, clearances, etc. It did a deep dive, provided tables and citations, etc. It was extremely convincing.
The problem? The numbers didn’t pass the gut check. I did a deeper dive, which took a few hours identifying ANSI standards and finding the correct information. Turns out, what ChatGPT recommended would have been against ANSI standards and extremely dangerous.
While it was clear in my circumstance, I’m sure there’s a lot of grey-er areas where it sounds convincing.
When it comes to Engineering, stick to your fundamentals. Don’t take AI’s information at face value. It can literally kill someone, significantly damage your company’s reputation, significantly hurt your career, etc.
[link] [comments]














English (US) ·