By the way, I forgot the name of the protagonist in Roses for Credit and asked AI. I wanted to see if it could do this task. AI told me the character’s name is Marilou, which it most certainly isn’t.
In questions relating to literature (which is 90%) of questions I ask AI, about half of responses I get is absolute crap. Mind you, I don’t ask AI to analyze works of literature. I only ask for straightforward facts. How many siblings does character such-and-such have and what are their names? When do the events in the novel take place?
The reason why I ask these questions is, for the most part, to be prepared for the weird things students will tell me about these texts in class.
Clarissa's Blog 
“about half of responses I get is absolute crap”
I don’t think that’s an accident, part of the purpose of AI seems to pretty clearly by to dumb down populations and wall off inconvenient or ‘dangerous’ facts.
LikeLiked by 1 person
I think more than half of the responses I get are crap. I have been asking a lot of questions about how to draw two routes on on Google API Map this past week. ALL the answers were crap, none of the coding that was submitted as an answer worked. I finally got a map to work but not from any AI help.
LikeLiked by 1 person
An Update:
I tried ChatGPT and got a good answer to my coding question. All my previous attempts were using Brave browser AI which was worthless. I’ll be going to ChatGPT in the future. All Ai’s are not created equally.
LikeLike
Funny, I had a similar experience last week. A friend was telling me about a book of stories she was reading. It sounded interesting so I wanted to recommend it to someone else. Instead of texting my friend, I asked chatgpt. I even included a really specific detail that should’ve made it easy to find. But no, it just hallucinated that detail in a bunch of unrelated books. So then I had to waste time fact-checking everything and realizing all the suggestions were wrong lol.
Not saying its totally useless. It saves me time on menial tasks and some technical coding stuff. But it’s not close to what it’s made out to be.
LikeLiked by 1 person
The flip side of the AI conversation is that it almost doesn’t matter whether AI actually works or not. Like the old saying goes, the market can stay irrational longer than you can stay solvent. Every company is pivoting to AI whether it makes sense for them or not, whether the tech delivers or not. A CTO today can’t afford not to go all-in on AI. Even if they’re skeptical, going against the grain means risking their job. If all your peers are shouting ‘AI is the future!’ and you say ‘uhh I’m not sure” you’re the loser and probably out of a job. It’s career insurance more than tech strategy at this point.
This is tulip-mania. I can smugly point out all the shortcomings and limitations but it won’t stop the stampede. They’re all still going to do it.
LikeLike
The real question is, can the internet stay functional longer than the AI can generate garbage?
If it means the death of the internet, it may turn out to be a net positive for humanity.
I’ve spent weeks repeatedly trying to contact a state govt-attached office to check on the status of a program I applied for. It uses some kind of customer service chatbot. I cannot get it to answer my question, it only spits out a form response to whatever keywords I’m using. I don’t know whether we qualified for the program or not, and can’t find out. And then I get an email follow-up. Same deal on the phone, but I spend ten minutes on hold, to talk to a real live person, who then *reads me the form response*.
This needs to die. Quickly.
LikeLiked by 1 person
The form resoonse is a step back in time to 3-5 yaers ago. If the state had brought in today’s generative AI, it would give you different, contradictory answers to the same question.
LikeLiked by 1 person