It’s been awhile. I hope all is well in the universe (other than the usual everything). I’ve been busy making content for classes and my clients that I’ll link to when I can! However, it is business time. Today I used ChatGPT to figure out how to replace a flange beneath a toilet, this is what I learned.
Generative A.I. only knows what words look like and how they interact with one another, their context is unfortunately lost in translation from pure logic seemingly input, with bias encoded unconsciously. Beyond that, it’s just wrong sometimes!
I had to correct multiple errors and missteps that the model had made, including the first section where the bot instructs us to simply use a towel and sponge for the quarter gallon of water we are about to dump on ourselves once we break the seal between the wax ring and the base of the porcelain (ceramic, pearl, gold plating, etc.). This immediately puts me on edge, I don’t think this is a smart way to start, nor is ignoring the entire conversation on the caps that belong on the toilet. I really don’t trust this software to deliver accurate and consistent expertise; Large language models seem to run on the saying, you are what you eat, and sometimes that’s poor information. If I looked at a source and it said A.I. generated, I would increase my skepticism further. I do think this is good for the world though; Digital citizens cannot rely on Google for their browsing, or at least, more people understand that now. It may seem bleak, but this chat bot misinformation can lead us towards media literacy programs and build a foundation of critical information gathering in us and our successors. I hope this inquiry is hopeful for you, it certainly is for me.
-Sincerely,
Pandora (Peyton)
Leave a Reply