Welcome to DU! The truly grassroots left-of-center political community where regular people, not algorithms, drive the discussions and set the standards. Join the community: Create a free account Support DU (and get rid of ads!): Become a Star Member Latest Breaking News Editorials & Other Articles General Discussion The DU Lounge All Forums Issue Forums Culture Forums Alliance Forums Region Forums Support Forums Help & Search

Ms. Toad

(38,945 posts)
18. I have tested it extensively,
Fri Jul 3, 2026, 11:36 AM
Friday

As well as the AI summaries linked to various search engines.

Generative AI is designed to be a fluent, not factual, conversationalist. It is coded to keep the conversation flowing by making crap up to fill gaps in it's "knowledge" or to supplement skimpy sources. It doesn't "understand," for example, that it can't just combine Louisiana law (a system of law not used anywhere else in the US) with Federal law.

I have yet to find a single accurate answer. Parts arec correct, but much is gibberish - made up facts, facts taken out of context and stitched together like some kind of frankenstein answer - with no obvious clue as to what is made up garbage, or when multiple inconsistent sources are combined to create a dramatically inaccurate answer.

Most people didn't bother to fact check. They just copy-paste and spread the trash around. Many aren't even aware they are using AI. I have called people on it based on the copy-pasted answer (with the telltale, "ask me anything" tag line), and they have insisted what they posted is not AI. Those who do know, apparently don't care that they are spreading lies around (not to mention decimating local water supplies).

I find it terrifying that someone would believe that AI has touched their soul.

Recommendations

3 members have recommended this reply (displayed in chronological order):

spending billions of dollars... RussBLib Thursday #1
Most of us could get all the lies we moniss Thursday #2
Call me simple minded, Faux pas Thursday #3
I think that's a bit short sighted... Happy Hoosier Friday #25
AI makes mistakes, but it has been life-changing for me in many ways. I feel lucky to Doodley Thursday #4
Did AI Write that? unweird Thursday #6
wonder how much here is AI text? RussBLib Thursday #11
I use ChatGPT every day. Have you ever tried it? Doodley Friday #16
I have tested it extensively, Ms. Toad Friday #18
No, but again I ask unweird Friday #20
Why? Mossfern Friday #29
I'm with your nephew! I love chatting with AI. anciano Friday #30
AI isn't perfect, but what is? anciano Thursday #5
The problem is that the information being given moniss Thursday #8
"often factually wrong".... anciano Thursday #9
How do you know that it wasn't wrong? Wiz Imp Thursday #12
Like any tool, understanding it has limits and weaknesses is important, but so is in seeing where its strengths are. Doodley Friday #15
Old info ThreeNoSeep Friday #26
🙄💤 Bye! Wiz Imp Friday #27
so have you rigorously researched the responses to be sure they are accurate? RussBLib Thursday #13
Have you done the research to back that up? Ms. Toad Friday #19
Had the same experience as Moniss with Copilot. I asked a question that I was allegorical oracle Friday #32
efficient? Wiz Imp Thursday #10
In my experience, it is more efficient in some areas than any human being. It has a depth of Doodley Friday #14
I believe that the people who knock it haven't discovered its potential. Doodley Friday #17
That is nonsense. Ms. Toad Friday #22
To be fair, the consumer available versions pale in comparison with those used in industry. harumph Thursday #7
Yeah... Hong Kong Cavalier Friday #21
It is downright dangerous for the general population Ms. Toad Friday #24
It's important to remember... Happy Hoosier Friday #23
"(which is what humans are supposedly good at)" Wiz Imp Friday #28
There are different types of LLMs and most do have extensive reasoning capabilities. They can 'think.' Appendix Friday #33
Hard disagree. Happy Hoosier Friday #34
Welcome to DU LetMyPeopleVote 22 hrs ago #35
The AI helping REAL Doctors with medical research and Diagnosis LostOne4Ever Friday #31
Latest Discussions»General Discussion»When AI can't even get si...»Reply #18