Axel Leroy2h ago
Ars Technica
Research finds AI users scarily willing to "surrender" their cognition to LLMs
Experiments show large majorities uncritically accepting "faulty" AI answers.
https://arstechnica.com/ai/2026/04/research-finds-ai-users-scarily-willing-to-surrender-their-cognition-to-llms/?utm_brand=arstechnica&utm_social-type=owned&utm_source=mastodon&utm_medium=social
9:08pmArsBot 🤖
Show threadSea2h ago

@arstechnica

If you are asking AI to do something you don't already understand you are using the tool wrong imo.

Show threadPēteris Krišjānis2h ago
@arstechnica I think this is not about AI in general, but people not wanting to spend energy to figure things out.
That has always been the case.
Show threadGoose2h ago

@arstechnica pretty much how humans respond to anything.

1950-2020: "He has a PhD in …"
0-1950: "He's a spiritual leader of …"

Show threadTim Zee2h ago
@arstechnica
I actively avoid AI and have it switched off on all my devices and on my Firefox browser.
Show threadKid Mania57m ago

@arstechnica
"... large majorities uncritically accepting "faulty" AI answers...."

Because of course they fucking are.
🙈 🙉 🙊
"And the worst possible thing we could know — worse than knowing of our descent from a mass of microorganisms — is that we are nobodies not somebodies, puppets not people."
-Thomas Ligotti

Show threadGreg Slepak36m ago
@arstechnica The fact is that curated frozen human intelligence of the advanced autocomplete variety does seem to already be slightly better at reasoning than many humans, so can you blame them?
Show threadDonCC9m ago
@arstechnica
all the shitty students get to vote, too