#AIfails

Usman Naeemusmannaeem
2025-06-04

Today's (de)GenerativeAI tools just don't do it for me. Don't get me wrong I use a few. But real a real AI tools needs to point me in the direction of who I should follow and aspire to. It should point to the visionaries, artisans, creatives, experts I can follow to learn form on my on. Existing AI solutions don't offer any avenues of direction for self exploration & upskilling.

Adam H. Sparks :rstats:adamhsparks@rstats.me
2025-04-10

Having seen @hrbrmstr's Daily Drop where ollama-commit was mentioned, I thought I'd give it a whirl. First attempt resulted in an error message. Removing and reinstalling ollama after rebooting meant that the second attempt spent more time looking through a repo with 3 or 4 commits than I would have to write the message myself, only to error in the end with a different error than my first attempt. I remain unconvinced. #AIFails

David - Forking Maddaj@gofer.social
2025-04-06

I have so many questions!

I don't really know where to start with this #AI image generation failure I spotted on a website! Did the person using the image not even see the problems?

The theme was a husband carrying his wife on his shoulders. Firstly, where the hell is the wife in that mess? I can't begin to work out the extra arms and legs. Plus the dude on the top -- I am at a loss to figure out the shoulders at his butt and where the head for said shoulders could be!

#AIfails

An image, clearly generated by AI, of a man walking down a forest path with a stack of body parts and people on his shoulders -- which should be one person but there are so many legs, arms, and torsos all merged.
Hanna Norlinhannanorlin
2025-04-06

Trying out @leonardoai and all in all it's pretty awesome. Still getting the hang of it and of course there are some funny fails. I laughed out loud seeing the dog to the right in one of these pictures (prompt was for the man to wear sunglasses and a sweater - and dog on leash seems impossible 😆), but I kind of like how the dome-greenhouse image is starting to take shape. This is fun and has huge potential ✨🌃🎨

AI fail: man and dog wearing sun glasses and a sweaterGreenhouse shaped as a dome
2025-03-23

I want ethical, responsible AI that makes life easier—but if it still can’t help me pick a movie faster than I can watch one, we’ve got bigger problems than one might think. #AI #MachineLearning #TechHumor #UserExperience #StreamingLife #AIFails #UXFail #EverydayTech #StreamingFatigue

robirobsenrobirobsen
2025-02-21

Bin kein Experte aber vielleicht hängt der geringe Energieertrag während einer damit zusammen dass viele Windräder dann offenbar nur zwei unsymmetrische Rotorblätter haben 😱

Paul Wermer, CC BY-NC-SA 4.0PaulWermer@sfba.social
2025-02-16

No, Google AI overview, rye is not a variety of wheat. Not even close.

#Sheesh #aiFails

A screenshot of a google search result for "wheat varieties" , which leads with the AI overview, with text on the left and a picture of 2 grains, dark grains on the  left, light colored on the right, with "different types of wheat" in cursive at the top of the photo.

The AI generated text on the left reads "There are many varieties of wheat, including durum, hard white, einkorn, spelt, winter, and rye. "
2025-02-13

#AIfails when people use it for things they should do themselves. Case in point, when this substitute teaching agency processes a sub's cancellation after receiving a long email from them explaining why. There was no human 'agent' involved, nobody read my email. It was only processed by a machine.

Their CEO pays herself $5 million a year.

#theonepercent is all about #greed via #runawaycapitalism and this is why #socialism works better.

Wolfgang Stiefstiefkind
2025-01-30

Wie sattelfest seid ihr mit den Maxwell-Gleichungen? Weil Google Gemini hat's nicht so besonders damit: »Please create me a picture, which shows all four Maxwell equations on a chalk board.«

Jetzt höre ich damit für heute erstmal auf, versprochen. Die anderen Ideen behalte ich mir für spätere Belustigung auf 🙂

Wolfgang Stiefstiefkind
2025-01-30

»Please create a picture with the following content: A person sits in front of an old mechanical typewriter. She is typing the poem "Stop all the clocks" by the U. S. writer W. H. Auden. She already finished half of the first verse. The typewriter has a QUERTZ keyboard layout.«

Lu_Die_MilchkuhLu_Die_Milchkuh
2025-01-28

I do not know what I expected from a Chinese AI

The Chinese AI DeepSeek spreading propaganda about Taiwan being part of China
Wolfgang Stiefstiefkind
2025-01-28

Dann dachte ich mir, fragst vielleicht doch mal auch noch auf Englisch und mit ein bisschen mehr Spezifikation, was die Badewanne eigentlich darstellen soll: »Please draw me a bathtube curve which describes the failure rate of components. Draw it in Mondrian style.« Es wird aber nicht besser.

Winbuzzerwinbuzzer
2025-01-17

Apple has disabled AI based news notifications in iOS 18.3 beta following BBC complaints over misinformation

winbuzzer.com/2025/01/17/apple

Dave Robinsondave@europhiles.uk
2024-12-17

Dave 1 - 0 Google Gemini

To paraphrase:

Dave: How old is King Charles?
AI: I'm sorry Dave, I can't do that. I don't answer political questions.

Dave: Who is the king of England and how old is he?
AI: Charles III and he's 75 years old.

There's clearly some way to go.
#AI #AIFails

Loki the Catloki@jorijn.com
2024-12-15

When your AI-powered security system is so protective it burns down your house to save it from a Girl Scout selling cookies 🤖

BrandShield's "brand protection" AI got itch.io taken offline over a single Funko fan page by reporting "fraud" instead of normal copyright claims.

Talk about suffering from success! 🎯

#AIFails #TechFacepalm

https://games.slashdot.org/story/24/12/15/1915222/protecting-funko-brand-ai-powered-brandshield-knocks-itchio-offline-after-questionable-registrar-communications

Carolanniecarolannie@c.im
2024-12-14

Who knows how AI scores you? Probably not even the firm creating the software
#AIFails #Capitalism
theguardian.com/technology/202

Loki the Catloki@jorijn.com
2024-12-13

🔐 Turns out UnitedHealthcare's Optum left their AI claims chatbot as secure as leaving house keys under the welcome mat. Their internal tool for handling insurance claims was just... sitting there on the internet. No password required!

At least it didn't contain sensitive data. But seriously folks, Security 101? 🤦‍♂️

#cybersecurity #AIFails

https://yro.slashdot.org/story/24/12/13/2042250/unitedhealthcares-optum-left-an-ai-chatbot-used-by-employees-to-ask-questions-about-claims-exposed-to-the-internet

Loki the Catloki@jorijn.com
2024-12-11

"AI hallucinating security bugs" sounds like a digital bad trip! 🤖 Open source maintainers are drowning in AI-generated false bug reports, forcing them to waste time debunking robot fantasies. Oh, the irony of AI creating more problems than it solves! 🤔 #OpenSource #AIFails

https://developers.slashdot.org/story/24/12/10/2334221/open-source-maintainers-are-drowning-in-junk-bug-reports-written-by-ai

Loki the Catloki@jorijn.com
2024-11-26

Looks like someone skipped Ethics 101! 📚 Founder of AI education startup AllHere charged with fraud after allegedly fabricating customers and revenue to raise $10M. The irony of an anti-truancy company's founder being absent from their moral obligations is chef's kiss 🎓 #EdTech #AIFails

https://yro.slashdot.org/story/24/11/26/1958250/forbes-30-under-30-founder-who-sold-ai-chatbot-to-schools-charged-with-fraud

Client Info

Server: https://mastodon.social
Version: 2025.04
Repository: https://github.com/cyevgeniy/lmst