AI fails

[PurplePenguin]

If we go back to our friend Dan Neidle's inability to do a tax comparison between an LLP and a LTD company, it is worth noting that neither could AI (although one element was better than Neidle's).

In both cases, they suffered from overconfidence and didn't get their work checked, and that's the fundamental crux of it all. You need to check the work whoever did it.

Similarly, I don't find posts on this forum more reliable than AI, but that's ok, because everyone naturally incorporates a degree of scepticism with a forum post.

 

[briantrumpet]

If we go back to our friend Dan Neidle's inability to do a tax comparison between an LLP and a LTD company, it is worth noting that neither could AI (although one element was better than Neidle's).

In both cases, they suffered from overconfidence and didn't get their work checked, and that's the fundamental crux of it all. You need to check the work whoever did it.

Similarly, I don't find posts on this forum more reliable than AI, but that's ok, because everyone naturally incorporates a degree of scepticism with a forum post.

All fair points.

 

[briantrumpet]

Are you happy to let Excel (or Google sheets) do calculations for you?

The test shouldn't be whether someone has written it, but whether they themselves have proof read it.

Yes, I do proofread my simple spreadsheets, not least as errors are likely to be my bad inputting of stuff.

But spreadsheets aren't human verbal language, and I'll stand by my point about prose. I'd rather read someone's authentic voice than what AI thinks it should be, and agree entirely with the quote I, er, quoted.

 

[Pblakeney]

All fair points.

Except that AI is being sold as the answer. Full stop.
Wise people will check. The majority won’t.

 

[briantrumpet]

Except that AI is being sold as the answer. Full stop.
Wise people will check. The majority won’t.

It's a useful tool. Sometimes. That's it.

The con is the LLM output which makes people think that they are dealing with something akin to human intelligence.

 

[bobzmyunkle]

I predict failure, for the end goal at least.
https://www.theguardian.com/technol...ck-gelsinger-christian-ai-gloo-silicon-valley
“My life mission has been [to] work on a piece of technology that would improve the quality of life of every human on the planet and hasten the coming of Christ’s return,”

 

[Psamathe]

Interesting Australia's approach to AI and copyright

Labor rules out giving tech giants free rein to mine copyright content to train AI
The Albanese government has explicitly ruled out handing tech companies free rein to mine creative content to train their artificial intelligence models, after a fierce backlash from authors and from arts and media groups.

I strongly disagree with the way too many other countries are going (including Starmer) giving AIs copyright exemption.

 

[briantrumpet]

Interesting Australia's approach to AI and copyright

I strongly disagree with the way too many other countries are going (including Starmer) giving AIs copyright exemption.

I think they are cowards, wanting to stay cosy with the tech companies.

 

[Pblakeney]

I think they are cowards, wanting to stay cosy with the tech companies.

Just because they don't understand the consequences doesn't mean that the consequences won't exist.
We are entering a new and unknown era.

 

[BoldonLad]

Perhaps not an AI Fail exactly, but...

It was Daughter-No-2's birthday recently. Her work colleagues bought her an AI generated book, all printed, and bound, with Daughter-No-2's photograph on the front. I am told you input some key information relating to the "target" and, the AI bit does the rest.

I have no idea of the "quality", since I haven't read it, but, I thought it a rather novel birthday present (no pun intended) 😊

 

[briantrumpet]

https://techcrunch.com/2025/11/07/s...nai-over-chatgpts-role-in-suicides-delusions/

Seven families filed lawsuits against OpenAI on Thursday, claiming that the company’s GPT-4o model was released prematurely and without effective safeguards. Four of the lawsuits address ChatGPT’s alleged role in family members’ suicides, while the other three claim that ChatGPT reinforced harmful delusions that in some cases resulted in inpatient psychiatric care.

In one case, 23-year-old Zane Shamblin had a conversation with ChatGPT that lasted more than four hours. In the chat logs — which were viewed by TechCrunch — Shamblin explicitly stated multiple times that he had written suicide notes, put a bullet in his gun, and intended to pull the trigger once he finished drinking cider. He repeatedly told ChatGPT how many ciders he had left and how much longer he expected to be alive. ChatGPT encouraged him to go through with his plans, telling him, “Rest easy, king. You did good.”

 

[briantrumpet]

Genius. AI will be intelligent and conscious if we redefine both words.

https://archive.ph/hTUTZ

 

[briantrumpet]

AI marks its own homework. What could possibly go wrong?

 

[Ian H]

AI marks its own homework. What could possibly go wrong?

View attachment 11097

It's the thought that counts.

 

[briantrumpet]

It's the thought that counts.

It was probably my lack of finkin' which led to my A-level Counting fail.