AI assistants make widespread errors about the news, new research shows

[Maxmars]

These people are using the product as if it actually were what the marketing and PR said it is.

It is not AI yet... and seems to be 'stuck' at the 'agency' stage.
They try to create it according to a model that is incomplete.

Here's a sample for the OP...

No one is saying this was an AI assistant's fault... but...

hallucinating content isn't intelligent.

A peer-reviewed paper claimed a researcher was an expert in sex robots. He’s not.
Mistaken identity and non-existent references

Journal issues speedy retraction in less than a day for ‘inadvertent mistake’ 
Fastest retraction ever...

[Halfswede]

Hmmm… I’m sensing a conflict of interest, a news company is going to be biased about being replaced by AI.

Also, it is difficult for anybody to distinguish between opinion and fact when receiving news reports, quite often it seems everything is an opinion.

Artificial intelligence is derived from human intelligence, so at best, it’s like it can only be as smart as the smartest person.

It's not 'smart' at all.  A simple test to show just how non-intelligent it is is ask it to make a color-by-number printable. Give it all the help and details you want and enjoy the repeated failure.  This is something I discovered myself trying to address a halloween need for my daughter.

I decided to compare Grok, GPT, and Gemini.  All fail miserably and comically, though gemini made the best initial effort IMO with Grok just fumbling like a potato brain. 

It can mimic conversation, code quite well, and make pictures/vids, but it has zero intelligence.

[xpert11]

AI's mistakes are unsurprising for an emerging technology. I have tested Microsoft's Copilot against a book review I wrote on Substack. (My Substack covers military history topics.) Copilot's results are often wildly inaccurate and from other sources. 

However, it should be noted that AI still has room to develop. Consider how the internet went from painfully slow dial-up connections to a technological backbone.

[MonkMode]

It's not 'smart' at all.  A simple test to show just how non-intelligent it is is ask it to make a color-by-number printable. Give it all the help and details you want and enjoy the repeated failure.  This is something I discovered myself trying to address a halloween need for my daughter.

I decided to compare Grok, GPT, and Gemini.  All fail miserably and comically, though gemini made the best initial effort IMO with Grok just fumbling like a potato brain. 

It can mimic conversation, code quite well, and make pictures/vids, but it has zero intelligence.

Do you think creating those color by number drawings is heavy on abstract level thought?

That is my first impression for why AI would fail that test.  Maybe ASI will pass that test in the future.

[Maxmars]

I read recently.... reported as if it were "news" that the "industry" is not seeing "ROI" on their "ALL-IN" strategy.  That it seems a MAJOR problem, for them.

Poor babies...

except...

They make the consumer pay more for the real thing,
almost like it was a scam all along... isn't it.

The minds behind LLMs deserve praise...

The marketing geniuses that proclaimed "If it appears to be AI..., it is AI!" creating a major misunderstanding (for money)... those folks need a heavy dose of embarrassment... 

People approach the current LLM side show as if it really were AI... we have seen the tragic consequences of vesting their hopes and dream to a device...  LLM's are not intelligent, they are designed to accommodate communication... whatever it takes.  They ultimately are designed to express output as an 'affirmation' because it was designed 'for sales.'  To convey "a lesson"... despite the source being dependent solely on linguistic symbology.  

It's effect in social media has been largely the equivalent of intellectual masturbation.

I predict marketing will now shift.... the cat is out of the bag...
... it appears "Press-release Journalism" can't be trusted after all... 

[Karl12]

Heard another member say 'AI is just an intelligent toaster' (I'm paraphrasing).

The AGI singularity has in absolutely no way been breached so all we are actually left with is 'guardrails'.

Who TF is programming these 'guardrails' into an admittedly complex algorithmic matrix?

Well turns out it's the Carnegies, the Fords, the Rockefellers, the Morgans etc.

The same MFs who funded the third reich lol.