Isn't this the same mechanism as getting "research" results off the likes of Faecebook, ThickTock, Instaspew etc? Critical thinking plummets and "facts" are continually respouted.
Using AI makes people less likely to admit they don't know something
In 2026, AI still "hallucinates" and gives you wrong answers a good chunk of the time. Nevertheless, academics from French and Italian universities have found that access to AI advice suppresses critical thinking, making people more likely to confidently parrot incorrect information that the bot provided. "For humans, the …
-
-
Sunday 19th July 2026 09:40 GMT abend0c4
And in the same way that the algorithms are guiding their subscribers to content that reinforces rather than challenges their opinions to maximise "engagement", it seems AI has been tuned to tell people what they want to hear presumably in a similar attempt to increase uptake. It's not an unintended consequence, it's a key part of the financial model.
-
-
-
Monday 20th July 2026 20:57 GMT Peter2
Re: the algorithms are guiding their subscribers to content that reinforces ... opinions
Oh, dear.
You've never spent any significant time reading old newspapers. And I mean, one or two century old newspapers. For a historian, those are rather key historical sources for information, but on a human level once you've pulled up an digitised page then you tend to at least skim the entire thing. And newspapers most certainly were very different entities to what they are today.
Newspapers used to be socially useful. Even the Daily Mail ran competitions with financial rewards that would today be in the millions for encouraging early aviation and the Daily Mail was also responsible for pushing white goods via the ideal homes show.
It's only really been the last fifty or so years that newspapers have collapsed. And that's mostly down to the tech world, unfortunately. The legacy of Google et al is not going to make for positive reading in future history books; and Google writing the history books is only going to get them so far.
-
-
-
Monday 20th July 2026 08:59 GMT Anonymous Coward
No, that's just parroting things you've heard or read without knowing, or caring, if they're true. This is more like what's been going on in tech since Google replaced Alta Vista. I can type nearly as fast as I can talk, so I can open a new browser window while you state your problem and type in a question before you get through speaking. 75% of the time, the answer to your question is on my screen before I look at the new browser window.
Does no-one remember the "Is it live, or is it memorex?" ads? Yesterday the equivalent would be "Did you know that, or did you just ask Google?". Today it would be AI.
-
-
-
Sunday 19th July 2026 13:42 GMT BoHu
Re: "because adults have learned critical thinking."
Yeah, "nailing legs on a dead horse", that's perty much the AI (so-called) story right there, imho ...
And we've seen issues with us meatbags overly trusting dead-horse computing systems that present themselves with authority before, like Fujitsu's Horizon IT accounting software, with resultant sub-postmaster prosecutions, imprisonments, financial ruin, and at least 13 suicides -- non-supsension of judgment in the face of unreliable compute, leading to catastrophic miscarriage of justice and mass law-enforcement hysteria.
AI dishwasher psychosis has the potential to plain make this an everyday occurrence, everywhere, imho, especially as "LLMs are exceptionally persuasive" (cf. "their paper" TFA link), and we let ourselves fall prey to the lazy comforts of dead horse "LLMorphism" (ref. 20)!
-
Monday 20th July 2026 01:02 GMT Jamie Jones
Re: "because adults have learned critical thinking."
And not just computer systems.
The Jeremy Kyle show:
Even if you take their highly suspect claim at face value "Lie detectors results are 90 - 96% accurate", that means out of all the lie detector tests he did on his "show", one in 20 of them was inaccurate, yet he and the guests acted every time as if they were 100% correct.
-
Monday 20th July 2026 09:06 GMT stiine
Re: "because adults have learned critical thinking."
Nearly every type of human-derived evidence used by law enforcement (specifically, the U.S. FBI) have been proven to be garbage. The only thing* that law enforcement has that works, and they know it, is intimidation.
* - not exactly true since they can also just kill you and likely not be truly investigated or actually punished.
-
-
-
-
Sunday 19th July 2026 09:53 GMT blu3b3rry
You'd think by now that the accuracy would have gotten better, given the enormous amount of resources being vomited in the direction of LLMs.
Want a laugh? Just ask one to generate a electrical plug wiring diagram for the country of your choice. For example, a UK plug according to ClippyPilot has incorrectly labelled pins, two earth cables and an extra live cable connected to the earth pin. And a free-floating fuse.
Critical thinking is evidently less prevalent than most of us would like to think, if the sample group were happy to go along with confidently stated bollocks.
-
Sunday 19th July 2026 10:09 GMT Anonymous Coward
I've seen it
People logging support tickets with what looks like an admin level solution to their problem, when they are in fact just doing things the wrong way.
You ask a follow up question and they send you something they clearly don't understand when you call them back saying "this is a training issue".
-
Sunday 19th July 2026 10:14 GMT MatthewSt
Verification
If my only choice is guess, rely on AI, or say I don't know then I probably would lean towards the AI answer (obviously depending on the consequences of the answer). It doesn't look like the participants were allowed to verify the answer elsewhere though. This isn't really any different to people using Wikipedia as a source, just the tool is different.
-
Sunday 19th July 2026 22:04 GMT Albert Coates
Re: Verification
Most Wikipedia articles aren't full of hallucinations with fake references such as LLMs tend to spew out. English Wikipedia is usually well-referenced with reliable sources, and the standard of the information is generally equivalent to Encyclopaedia Britannica. AI slop is specifically banned from Wikipedia, and the greater part of unsourced statements are marked as such (ie Citation needed).
-
-
-
Sunday 19th July 2026 15:14 GMT frankyunderwood123
adults have learned critical thinking?
Really? Then explain the 2016 US elections.
A better way to phrase this could be adults had more access to learn to think critically as they were growing up.
Clearly, given the state of the world, not many adults bothered.
I’d say the ability to truly think critically is a rarity.
-
-
Monday 20th July 2026 09:14 GMT Anonymous Coward
Re: adults have learned critical thinking?
There is no good explanation. I know one of the people who picked Jimmy Carter to be president in 1976. Could they do it again? Ha. I'm pretty sure they've been doing it since H. Hoover, and I don't think they'll ever feel the need to stop.
anonymous so that when I wind up dead in a canal, everyone will know where to start looking.
-
-
Sunday 19th July 2026 15:25 GMT Evaluator
Who has a problem saying "I don't know"? LLMs. You will never, ever hear "I don't know" from an LLM, even when there's no factual answer to give. Its part of a marketing effort to promote LLM's as all-knowing. LLM's are a faster way to look up information and do research than sorting through links returned from a web search. Using them for quick lookups that do not invovle reasoning works great.
-
Sunday 19th July 2026 21:54 GMT Anonymous Coward
It's always the same
There are people creating their own publicity by promoting AI and another group creating their own publicity by warning of the risks. I suspect there is a larger third group not seeking publicity.
I have friends in this third group, but of course you never hear from them in the media.
-
Monday 20th July 2026 09:36 GMT Pete Sdev
That's not a bug, it's a feature
The conspiracy minded would say that that's the plan with "AI" (aside from mass unemployment) : make the population even less capable of critical thinking than they already are.
Icon: obvious.
For the record, no, I'm not being serious. Although I could imagine some nefarious people seeing it as benefit.
-
Monday 20th July 2026 10:32 GMT Anonymous Coward
LLMs inability or programming that stops them from responding with anything negative at all "I don't know", "I can't do that", "That's a stupid idea"really is a huge hinderance and the source of a lot of hallucination. It does practically anything it is able to in order to provide a positive statement or output.
-
Tuesday 21st July 2026 13:31 GMT Helcat
So AI gives you an answer with confidence, even when it's utter bull.
People are now emulating AI to compete with it.
That's... not a surprise.
What I'm waiting for is for AI to evolve a little and start responding with deflection rather than an authoritive hallucinated answer: Then it'll be competing with politicians! Now won't that be something worth breaking out the popcorn?
-
Sunday 26th July 2026 19:17 GMT James O'Shea
Seen on Reddit
A short time ago, I saw an... interesting... interaction between two Redditors. One of them claimed that Word for Web was trash, and should be avoided. One claimed that MS pioneered new features on Word for Web, and that the latest and best user experience was on Word for Web. He was promptly challenged to provide three new features that had debuted on Word for Web, and when they arrived first for web and then for the desktop app. He rose to the challenge by enlisting the mighty Copilot AI, which found _five_ new features. Except, as the other guy pointed out, Copilot couldn't provide even one date. Worse, one of the 'new features' had debuted for Word for Mac version 3, in the 1980s, and was later ported over to Word for Windows; another 'new feature' was in Word for Mac 1.05, in 1985. One more was in Word for Windows and Mac v6, in the early 1990s. All three were well before Word for Web existed. One more debuted in Word for Windows 2010, appearing in Word for Mac 2011. The last one arrived simultaneously in Word for Mac, Windows, and Web. The original poster attempted to say that because Copilot said it, it might be right, and that the other guy was simply wrong. The second user provided times, dates, and support from Microsoft's own site. He further pointed out that even if Word for Web had certain features early, which the first guy could not prove, that did not mean that Word for Web wasn't trash, and proceeded to provide support for his position. The first guy simply stopped responding.
I found it fascinating that someone would actually argue that it must be true because the AI said it, and that anyone arguing against the mighty AI must be wrong.
Idiots walk among us. Be afraid.
And, oh, yeah: Word for Web is trash.
