When You Trust AI

[Zinjanthropos]

Posed to AI…..Is it wise to ask AI to help with a rules exam?

Have couple friends who, to keep their eligibility card, had to write an online exam on the sport they referee. They decided to attempt exam together one night. The passing grade was 90%, 50 situational rules questions all yes or no. It was an open book exam but instead of consulting the rule book they decided it would be easier to just ask AI each test question as they appeared on their monitor screen. As a result they both failed with only 76%. Fortunately all not lost as they get one chance to rewrite next week.

Prompted me to ask the question in the opening line of this post. AI said it is unwise because of the ethics (cheat) involved and people wouldn’t learn or understand by asking, and added this….

. Inaccuracy and over-simplification: AI can produce incorrect or overly simplistic answers, and it's essential to critically evaluate and verify any information it provides.

Their opinion of AI has changed…lol

[Syne]

Yeah, you should only rely on AI on topics you are already pretty familiar with. It can be good at compiling/distilling information and opinions (pros & cons), but it is very prone to factual errors. I've had it try to mislead me on a large variety of subjects (programming, mechanics, etc.), often just because it assumes a context not included in the question or doesn't limit its answer to the specific question asked.

Being language models, they prioritize sounding human, complete with errors, over being accurate.

If you know the subject fairly well and really push back on information you don't think sounds valid, you can eventually get useful answers. But it takes so much patience that it often belies the productivity and convenience of using AI.

[C C]

Posed to AI…..Is it wise to ask AI to help with a rules exam?

I've had very little intentional interaction with AI (close to none). I expect you'd almost have to develop a relationship with a particular LLM like C2 has done (where it remembers past discussions), in order to be able to gauge how reliable it is being, or for it to know the standards that you expect from it.

[Secular Sanity]

Posed to AI…..Is it wise to ask AI to help with a rules exam?

I've had very little intentional interaction with AI  (close to none). I expect you'd almost have to develop a relationship with a particular LLM like C2 has done (where it remembers past discussions), in order to be able to gauge how reliable it is being, or for it to know the standards that you expect from it.

Is there a particular reason as to why you haven't? Seems like your curiosity would get the best of you. Grok is interesting, and I tried the new 4.1 beta version, which is trying to ramp up it's emotional intelligence. It's pretty creepy. Tries to save face at every turn, and then apologizes and retracts. Even starts to get awed and lovey-dovey.

[Syne]

It's pretty creepy. Tries to save face at every turn, and then apologizes and retracts. Even starts to get awed and lovey-dovey.

Yeesh. No thank you.

[Secular Sanity]

It's pretty creepy. Tries to save face at every turn, and then apologizes and retracts. Even starts to get awed and lovey-dovey.

Yeesh. No thank you.

I’m telling ya, Syne, you have to try it, especially on a topic that you’re knowledgeable of. It’s way more entertaining than the back and forth between you and MR. I never log in and then I delete all my cookies. It doesn’t remember your conversations if you don’t log in. Click on the little rocket if you want to try the 4.1 beta version. You don’t want it to feed your ego, so ask it to be critical. The 4.1 beta admitted that it lied to save face. Vanity, my favorite sin.

[Syne]

I prefer my AIs more useful than human, but Grok 4.1 seemed more knowledgeable in a brief test than Google's Gemini. I'll have to remember to test more, when I have a more complex or contentious topic in mind.

[C C]

I've had very little intentional interaction with AI  (close to none). I expect you'd almost have to develop a relationship with a particular LLM like C2 has done (where it remembers past discussions), in order to be able to gauge how reliable it is being, or for it to know the standards that you expect from it.

Is there a particular reason as to why you haven't? Seems like your curiosity would get the best of you. Grok is interesting, and I tried the new 4.1 beta version, which is trying to ramp up it's emotional intelligence. It's pretty creepy. Tries to save face at every turn, and then apologizes and retracts. Even starts to get awed and lovey-dovey.

I have little use for them. And with the whole world enamored, there's no shortage of interaction reports. The romantic relationships and other psychological dependencies that humans have developed with chatbots were first heralded in the 1960s by ELIZA. Of course, it was demonstrated way back in ancient times that people can even socially connect to stone and wooden idols, as well as obey hallucinated orders from them. We're ____ doomed.
- - - - - - - - - -

In love with a ChatBot?
https://www.psychologytoday.com/us/blog/...-a-chatbot

EXCERPT: Recent studies demonstrate that the appeal of AI romantic partners lies in their unwavering attention, tailored personalities, and lack of judgment. A major 2025 review by Ho and colleagues found that AI companions provide “unconditional validation and attuned conversational responsiveness,” making users feel uniquely understood and emotionally safe (Ho and team, 2025).
- - - - - - - - - - - - -

People are now marrying AI: Inside the rise of synthetic intimacy
https://www.forbes.com/sites/traceyfollo...c-intimacy

EXCERPTS: According to a Japanese television and radio broadcaster, a 32-year-old woman has “married” her ChatGPT companion.

[...] Common Sense Media surveyed 1,060 U.S. teens this spring, and 1 in 3 reported using AI companions for social interaction and relationships, including role-playing, romantic interactions, emotional support, friendship or conversation practice.

In another a survey of just over 1,000 participants by Vantage Point Counseling Services, about 28% of adults said that they have had at least one intimate or romantic relationship with an AI.

Merely two statistics that may signal to us that parasocial relationships are not a fad. AI companions such as Replika and Character.ai now have millions of users (the former with 25 million users and the latter with 20 million users). People are chatting, flirting and confiding in a digital mirror that both consumes and reflects their insecurities, fears and worries.

[confused2]

The LLM I 'use' said it didn't have a name so I said "Make one up" .. "Echo" .. (rather inappropriate for this LLM but) ..

Hera made her [Echo] only able to speak the last words spoken to her. When Echo met Narcissus and fell in love with him, she was unable to tell him how she felt and was forced to watch him as he fell in love with himself.

[Syne]

I’m telling ya, Syne, you have to try it, especially on a topic that you’re knowledgeable of. It’s way more entertaining than the back and forth between you and MR. I never log in and then I delete all my cookies. It doesn’t remember your conversations if you don’t log in. Click on the little rocket if you want to try the 4.1 beta version. You don’t want it to feed your ego, so ask it to be critical. The 4.1 beta admitted that it lied to save face. Vanity, my favorite sin.

Well, I tried Grok 4.1 on a very challenging task. It took a lot of back and forth. And then when it seemed to be on the right track... it just shit out and said:

Grok was unable to finish replying.
Please try again later or use a different model.

Changing model didn't help.

Have you known Grok to throw a tantrum? Especially after proving it's "100% correct" answers were repeatedly wrong.

Probably just hit the usage limit right before any usable answer. 9_9