#hallucination — Public Fediverse posts

A friend butt-dialed me at 01:00 last night. It diverted to voice mail, which recorded a minute of muffled background noises. T-Mobile's amazing "A.I." then transcribed its audio hallucination in Hiragana, which makes perfect sense because I was in Japan with my phone a little over 2 years ago. We live in an age of wonders!
#AI #TMobile #ButtDial #hallucination

A friend butt-dialed me at 01:00 last night. It diverted to voice mail, which recorded a minute of muffled background noises. T-Mobile's amazing "A.I." then transcribed its audio hallucination in Hiragana, which makes perfect sense because I was in Japan with my phone a little over 2 years ago. We live in an age of wonders!
#AI #TMobile #ButtDial #hallucination

From hallucination to verification www.pixelsandpatterns.org/p/from-halluci… #AI #citations #hallucination

From hallucination to verification www.pixelsandpatterns.org/p/from-halluci… #AI #citations #hallucination

From hallucination to verification www.pixelsandpatterns.org/p/from-halluci… #AI #citations #hallucination

From hallucination to verification www.pixelsandpatterns.org/p/from-halluci… #AI #citations #hallucination

From hallucination to verification www.pixelsandpatterns.org/p/from-halluci… #AI #citations #hallucination

RE: https://floss.social/@ramblingsteve/116945912010965459

"So basically people became much worse – the accuracy was only one third – but they were twice as confident," he said.

…

"I'm very much concerned for children, because adults have learned critical thinking. But for children who basically are born with these systems, the risk is that they don't even learn the basic critical skills."

#hallucination #confidence #LLM #CriticalThinking

RE: https://floss.social/@ramblingsteve/116945912010965459

"So basically people became much worse – the accuracy was only one third – but they were twice as confident," he said.

…

"I'm very much concerned for children, because adults have learned critical thinking. But for children who basically are born with these systems, the risk is that they don't even learn the basic critical skills."

#hallucination #confidence #LLM #CriticalThinking

RE: https://floss.social/@ramblingsteve/116945912010965459

"So basically people became much worse – the accuracy was only one third – but they were twice as confident," he said.

…

"I'm very much concerned for children, because adults have learned critical thinking. But for children who basically are born with these systems, the risk is that they don't even learn the basic critical skills."

#hallucination #confidence #LLM #CriticalThinking

RE: https://floss.social/@ramblingsteve/116945912010965459

"So basically people became much worse – the accuracy was only one third – but they were twice as confident," he said.

…

"I'm very much concerned for children, because adults have learned critical thinking. But for children who basically are born with these systems, the risk is that they don't even learn the basic critical skills."

#hallucination #confidence #LLM #CriticalThinking

RE: https://floss.social/@ramblingsteve/116945912010965459

"So basically people became much worse – the accuracy was only one third – but they were twice as confident," he said.

…

"I'm very much concerned for children, because adults have learned critical thinking. But for children who basically are born with these systems, the risk is that they don't even learn the basic critical skills."

#hallucination #confidence #LLM #CriticalThinking

@alainmi11 et quitte à faire le choix de l’IA (c’est tellement un bulldozer qu’on peut comprendre que des personnes pensent qu’on DOIT utiliser ce truc), pourquoi ne pas prendre 15 secondes pour évaluer le résultat avant de poster une merde ? #slop #noiag #hallucination

@alainmi11 et quitte à faire le choix de l’IA (c’est tellement un bulldozer qu’on peut comprendre que des personnes pensent qu’on DOIT utiliser ce truc), pourquoi ne pas prendre 15 secondes pour évaluer le résultat avant de poster une merde ? #slop #noiag #hallucination

@alainmi11 et quitte à faire le choix de l’IA (c’est tellement un bulldozer qu’on peut comprendre que des personnes pensent qu’on DOIT utiliser ce truc), pourquoi ne pas prendre 15 secondes pour évaluer le résultat avant de poster une merde ? #slop #noiag #hallucination

@alainmi11 et quitte à faire le choix de l’IA (c’est tellement un bulldozer qu’on peut comprendre que des personnes pensent qu’on DOIT utiliser ce truc), pourquoi ne pas prendre 15 secondes pour évaluer le résultat avant de poster une merde ? #slop #noiag #hallucination

@alainmi11 et quitte à faire le choix de l’IA (c’est tellement un bulldozer qu’on peut comprendre que des personnes pensent qu’on DOIT utiliser ce truc), pourquoi ne pas prendre 15 secondes pour évaluer le résultat avant de poster une merde ? #slop #noiag #hallucination

Here is another "invention" of mine #vibecoding. That inherently mitigates the #hallucination preponderence of #LLM

A pane from the Engine configuration screen.

I have a consensus panel assembly. A council of (local) LLMs. They look at a simple query (easy class) and if there is any substantive disagreement, they bounce it up to a higher model.

This way, if they get the dumbs, rather than it sneaking into the final output, it gets escalated to the bigger brainz.

Its vaguely based on the space shuttle consensu mainframe design. So I can not claim the credit for it.

HALLUCINATION (Estats Units) presenta nou àlbum: "Something Lost, Something Gained" #Hallucination #DeathMetal #Juliol2026 #EstatsUnits #NouÀlbum #Metall #Metal #MúsicaMetal #MetalMusic

https://www.europesays.com/ie/579487/ Cincinnati poison control expert shares concerns about social media ‘Benadryl challenge’ #BenadrylChallenge #cincinnati #Éire #hallucination #Health #Heart #HighDose #IE #increase #IntentionalBenadrylOverdose #Ireland #lot #Medication #Ohio #parent #People #PoisonControl #PoisonInformationCenter #Seizure #SocialMedia #SocialMedium #teen #yin

"Phantom squatting” uses #AI hallucinated #domains for #cyber attacks

source: cybernews.com/security/phantom…

This new #attack vector works as follows: criminals probe AI models by asking questions such as "What's the official #website for company X?" or where a particular file, application, or package can be downloaded. Sometimes, AI systems hallucinate even when responding to these seemingly simple questions.

What's more, Unit 42 found that models can repeatedly generate the same non-existent domains. Attackers can therefore collect these hallucinated domains, #register the available ones, and proceed with a familiar #cybercrime playbook: lure victims to #fake websites and serve #malware, distribute malicious packages, or steal credentials through #phishing attacks.

#news #cybersecurity #security #domain #hallucination #technology #software #internet #online #fail #problem #www #web #cyberattack #knowledge #browser #knowhow

"Phantom squatting” uses #AI hallucinated #domains for #cyber attacks

source: cybernews.com/security/phantom…

This new #attack vector works as follows: criminals probe AI models by asking questions such as "What's the official #website for company X?" or where a particular file, application, or package can be downloaded. Sometimes, AI systems hallucinate even when responding to these seemingly simple questions.

What's more, Unit 42 found that models can repeatedly generate the same non-existent domains. Attackers can therefore collect these hallucinated domains, #register the available ones, and proceed with a familiar #cybercrime playbook: lure victims to #fake websites and serve #malware, distribute malicious packages, or steal credentials through #phishing attacks.

#news #cybersecurity #security #domain #hallucination #technology #software #internet #online #fail #problem #www #web #cyberattack #knowledge #browser #knowhow

"Phantom squatting” uses #AI hallucinated #domains for #cyber attacks

source: cybernews.com/security/phantom…

This new #attack vector works as follows: criminals probe AI models by asking questions such as "What's the official #website for company X?" or where a particular file, application, or package can be downloaded. Sometimes, AI systems hallucinate even when responding to these seemingly simple questions.

What's more, Unit 42 found that models can repeatedly generate the same non-existent domains. Attackers can therefore collect these hallucinated domains, #register the available ones, and proceed with a familiar #cybercrime playbook: lure victims to #fake websites and serve #malware, distribute malicious packages, or steal credentials through #phishing attacks.

#news #cybersecurity #security #domain #hallucination #technology #software #internet #online #fail #problem #www #web #cyberattack #knowledge #browser #knowhow

Cincinnati poison control expert shares concerns about social media ‘Benadryl challenge’

A social media trend called the “Benadryl challenge” is raising concerns in…
#NewsBeep #News #US #USA #UnitedStates #UnitedStatesOfAmerica #Medication #benadrylchallenge #Cincinnati #hallucination #Health #heart #highdose #increase #intentionalbenadryloverdose #lot #Ohio #parent #people #poisoncontrol #poisoninformationcenter #Seizure #SocialMedia #socialmedium #teen #yin
https://www.newsbeep.com/us/753469/

Cincinnati poison control expert shares concerns about social media ‘Benadryl challenge’

A social media trend called the “Benadryl challenge” is raising concerns in…
#NewsBeep #News #US #USA #UnitedStates #UnitedStatesOfAmerica #Medication #benadrylchallenge #Cincinnati #hallucination #Health #heart #highdose #increase #intentionalbenadryloverdose #lot #Ohio #parent #people #poisoncontrol #poisoninformationcenter #Seizure #SocialMedia #socialmedium #teen #yin
https://www.newsbeep.com/us/753469/

HalluSquatting turns AI hallucinations into an automated botnet-building weapon

https://1ban.news/hallusquatting-ai-hallucination-agentic-exploit-2026/
#1ban #hallusquatting #hallucination #agentic #exploit #tech

Claude says it can't resolve a FQDN to an IP address, because robots.txt forbids it.

#claudeSonnet5 #hallucination

Claude says it can't resolve a FQDN to an IP address, because robots.txt forbids it.

#claudeSonnet5 #hallucination

Somebody had the gall to submit AI-generated, hallucinated documents to the Supreme Court. And the 'methyl isocyanide' comment about AI is hilarious. Imma start calling toxic people methyl isocyanide from now. Methyl isocyanide is an AI hallucinogen. 🤣 https://www.ndtv.com/india-news/invisible-insidious-top-court-warns-against-use-of-ai-in-legal-process-11716803

#AI #hallucination

Somebody had the gall to submit AI-generated, hallucinated documents to the Supreme Court. And the 'methyl isocyanide' comment about AI is hilarious. Imma start calling toxic people methyl isocyanide from now. Methyl isocyanide is an AI hallucinogen. 🤣 https://www.ndtv.com/india-news/invisible-insidious-top-court-warns-against-use-of-ai-in-legal-process-11716803

#AI #hallucination

Here is some early AI for you…

Way back in 2012, while doing research for the very first version of my AI lecture I was testing an AI chatbot app. Every day it would send me a check-in notification.

The communication was generally benign. But most of it was anything but intelligent. I captured this gem one morning. It’s pretty indicative of the “discussions” I had with it.

This is the talk I gave in Romania, shortly after this chatbot lost its damn mind:

https://vid.northbound.online/w/5VZnJx6rnCt5VFmfN9r6jr

#AI #ArtificialIntelligence #ChatBot #Hallucination #OldSchool #Technology

Here is some early AI for you…

Way back in 2012, while doing research for the very first version of my AI lecture I was testing an AI chatbot app. Every day it would send me a check-in notification.

The communication was generally benign. But most of it was anything but intelligent. I captured this gem one morning. It’s pretty indicative of the “discussions” I had with it.

This is the talk I gave in Romania, shortly after this chatbot lost its damn mind:

https://vid.northbound.online/w/5VZnJx6rnCt5VFmfN9r6jr

#AI #ArtificialIntelligence #ChatBot #Hallucination #OldSchool #Technology

Just read a research proposal which referenced the ECJ judgment in C-413/23 P EDPS v SRB, and spelled out the latter acronym as “Scarlett Ring Bearer”. That’s not just an #AI #hallucination but the beginning of AI fairy tale! Who can do better? #aislop

Just read a research proposal which referenced the ECJ judgment in C-413/23 P EDPS v SRB, and spelled out the latter acronym as “Scarlett Ring Bearer”. That’s not just an #AI #hallucination but the beginning of AI fairy tale! Who can do better? #aislop

Google hallucinated that I am sponsored by Ground News

https://www.youtube.com/post/UgkxsAIUDMQTYc15yQqURsGy0Kg7O1x2TtnB

#HackerNews #Google #GroundNews #Hallucination #TechNews #AI

Google hallucinated that I am sponsored by Ground News

https://www.youtube.com/post/UgkxsAIUDMQTYc15yQqURsGy0Kg7O1x2TtnB

#HackerNews #Google #GroundNews #Hallucination #TechNews #AI

🤖🥳 Oh joy, another #AI model that can't stop hallucinating! GPT-5.5 is having a triple #hallucination party while GLM-5.2 awkwardly watches from the corner. 👀 But hey, bigger is always better, right? Just ask the banned models - they have all the free time to ponder that profound wisdom. 🧠🚫
https://arrowtsx.dev/bigger-models/ #GPT5.5 #GLM5.2 #TechNews #AIModels #HackerNews #ngated

🤖🥳 Oh joy, another #AI model that can't stop hallucinating! GPT-5.5 is having a triple #hallucination party while GLM-5.2 awkwardly watches from the corner. 👀 But hey, bigger is always better, right? Just ask the banned models - they have all the free time to ponder that profound wisdom. 🧠🚫
https://arrowtsx.dev/bigger-models/ #GPT5.5 #GLM5.2 #TechNews #AIModels #HackerNews #ngated

KPMG report contained AI hallucinations on benefits of . . . AI

Saving this here for next year’s teaching:

A KPMG report on how AI is being used by businesses across the world exaggerated adoption of the technology with bogus case studies that appear to have been based on AI hallucinations. The October report, “Redefining excellence in the age of agentic AI”, made numerous false claims about the use of AI by organisations including the Swiss bank UBS, the UK’s National Health Service and the public transit groups Swiss Federal Railways and Transport for London. The inaccuracies were identified as AI hallucinations by the research group GPTZero and verified by the FT. After being alerted to the issue, UBS said it would ask KPMG to remove the false claims, and the Big Four firm on Thursday pulled the report from some of its websites. The discovery is the latest in a number of apparent AI hallucinations in reports by professional services firms and follows EY’s retraction of a study last month over fake footnotes and other errors identified by GPTZero.

https://www.ft.com/content/b3828e92-4961-4b39-84f0-c42f33be3c3f?syn-25a6b1a6=1

KPMG report contained AI hallucinations on benefits of . . . AI

Saving this here for next year’s teaching:

A KPMG report on how AI is being used by businesses across the world exaggerated adoption of the technology with bogus case studies that appear to have been based on AI hallucinations. The October report, “Redefining excellence in the age of agentic AI”, made numerous false claims about the use of AI by organisations including the Swiss bank UBS, the UK’s National Health Service and the public transit groups Swiss Federal Railways and Transport for London. The inaccuracies were identified as AI hallucinations by the research group GPTZero and verified by the FT. After being alerted to the issue, UBS said it would ask KPMG to remove the false claims, and the Big Four firm on Thursday pulled the report from some of its websites. The discovery is the latest in a number of apparent AI hallucinations in reports by professional services firms and follows EY’s retraction of a study last month over fake footnotes and other errors identified by GPTZero.

https://www.ft.com/content/b3828e92-4961-4b39-84f0-c42f33be3c3f?syn-25a6b1a6=1