DEV Community

Khalifah Shabazz
Khalifah Shabazz

Posted on

AI Deleting Trust

AI Deleting Trust

AI is most assuredly deleting any kind of trust people did have in tech, what little there may have been. Today I'm looking at an article and AI (at least I think it was AI) got the name wrong of a missing person.

I opened my start menu on Windows 11 today and saw an image with a headline about a missing person update. So I clicked to find out more.

Screenshot taken from Windows 11 start menu about a missing person update

Immediately I'm taken to MSN where it shows me several sections about the missing person. I noticed how the biggest section had the name wrong. The name should be as it is on the right of the screenshot.

Screenshot of MSN page showing several content sections about the missing person

When you click the link to go to the full story, the incorrect name of the
person persists.

Screenshot of the MSN page about the missing person and where the incorrect name of the missing person persists

However, if I go to where MSN ripped this story from, the name is correct.

Screenshot of the ABC new page where MSN may have copied the article from, but the name is correct

I know that as we dig deeper into a world with AI, we can clearly see what it is good at and clearly what it is not. My current evaluation is that its fine to use AI for anything that is not fact based. Here's my opinionated list, based on my experience, of which I may use AI.

Good uses:
Help Write/Update your resume: Yes
Cheap, quick design work: When it doesn't matter, yes!
Fictional writing: Absolutely!
Rubber duck programming: Yes.
Helping it profile you: YES!
Anything else: NO!

Bad Uses:
Fact checking: NO!
Coding: ABSOLUTELY NOT!
Good Original Design work: ABSOLUTELY NOT!
Exploring concepts: You better be careful, this only works if you're already an expert on the subject.
Learning: No

That last one needs some explanation. In the recent months I tried to use AI to help with learning several subjects. For the sake of this article I'll use my experience using AI to help learn more about Kubernetes. When I would get stuck I would turn to AI to see if I could get back on track quickly. I had a particularly rough time setting up a Network CNI addon that I could use to practice network policies in a cluster. I was already having a hard enough time with the partial documentation the plugins have, but adding AI to the mix made things take longer to solve. As I realized that AI was wasting my time because it was giving me half-baked solutions. AI does NOT test anything it gives you. I'm not sure even if AI learned anything from that experience, which would mean I completely wasted both our time and energy that AI needs to run.

From now on I'm learning the old-fashioned way. I don't use AI much since it seems to be a glorified search engine that talks back under the pretense of a human, and uses a lot of energy. It makes me wonder about stories I hear about researchers and scientist using AI. Are they really or is that a load of marketing hogwash?

Top comments (0)