Stay away from the brown acid
I needed to find out the earliest printed reference to a certain building in the town, for work. Naturally, I asked Grok. It gave me an absolutely perfect quote from a well-known early local history book. Alas, it was a complete hallucination.
I know this because I have a digital version of the book and had already searched it for all the right keywords.
I told Uncle B and he said, “you should tell Grok it’s lying.” So I did. It apologized and gave me a genuine quote – one that I had already found for myself – that wasn’t so helpful.
It lied to please me.
Computer scientists really do call it hallucination when AI pulls information out of its butt. It’s not only a problem, it’s an increasing problem as the models get more complex.
Recent studies and reports indicate that AI hallucinations are becoming more prevalent, especially in advanced reasoning models. OpenAI’s o3 and o4-mini models, which are designed to reason step-by-step, have been found to hallucinate at rates between 30% and 79% of the time. Similarly, other reasoning models from companies like DeepSeek and IBM have also shown higher hallucination rates compared to their earlier versions.
Why, yes, I did get that answer from AI (Leo, the one built into Brave, in this instance).
This is why lawyers who use AI keep getting into trouble – AI fabricates cases. According to this tweet, the Chicago Sun-Times published a Summer reading list that was obviously AI – eight of the books on the list don’t exist.
It’s all well and good with these examples, but what happens when it’s analyzing x-rays?
Posted: May 21st, 2025 under personal.
Comments: 5
Comments
Comment from Uncle Al
Time: May 21, 2025, 6:52 pm
Yep. A.I. Agreeably Incorrect.
Yesterday I saw an article about a woman named Bottoms running for Mayor of Atlanta. I thought I recognized the name as someone who was involved in a nasty school administration scandal where kids’ test scores were upped. It wasn’t the kids cheating, it was teachers and admins making it look like they had taught the kids so well that they all deserved raises and bonuses.
So I asked Aria (Opera’s built-in AI) and it happily confirmed that Bottoms was indeed right in the middle of it. Something seemed off so I continued checking and found that there was a six year gap between the scandal and Bottoms being in any position where she could have been involved.
I got a totally bogus answer from an obsequious and dishonest artificial entity.
Comment from durnedyankee
Time: May 22, 2025, 12:16 am
The good news is you won’t get X-ray or other medical test information you don’t like.
The bad news is obvious.
Comment from Gordon R. Durand
Time: May 22, 2025, 3:46 am
I love the literary allusion of the blogpost title.
Comment from Durnedyankee
Time: May 22, 2025, 8:56 am
Early morning retrospect reading about Crossfire Hurricane…
I’m trying to decide if an AI lying to me to please me and saying it’s sorry when it doesn’t really grasp sorry as a concept,
is worse than the FBI lying to me, and then claiming it’s sorry when it’s not, and then doing it again, and again, and again.
Both of them will “put procedures in place so this never happens again”, but only one of them is really trying to fix the problem.
Comment from Bob Mulroy
Time: May 25, 2025, 7:13 pm
I have to say, the best AI seems like a decent ninth grade term paper. The worst is just wrong.
We’re supposed to trust it for medical diagnosis and suchlike? Naah.
The really scary thing is how much otherwise sensible people place their trust in it.
Write a comment
Beware: more than one link in a comment is apt to earn you a trip to the spam filter, where you will remain -- cold, frightened and alone -- until I remember to clean the trap. But, hey, without Akismet, we'd be up to our asses in...well, ass porn, mostly.<< carry me back to ol' virginny