#hallucinations — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #hallucinations, aggregated by home.social.
-
#AI #AIslop #hallucinations
1) ↙️French #EODev_H2 GEH2 hydrogen fuel cell introduced in Australia↘️
2) Yes, #energy_observer had a Proton #nelhydrogen electrolyzer; GEH2 only has a 70kW #Toyota fuel cell.
3) No electrolyzer, so no metal hydride storage tanks for #H2 storage either. (uses tanks of H2).
#fuelcell #hydrogen -
Почему RAG — фундамент любой AI-трансформации
За последние годы большинство AI-проектов в компаниях стартуют одинаково: сначала делают чат-бота, затем добавляют агентов, автоматизируют отдельные процессы и ожидают роста эффективности. На практике такие проекты часто не дают устойчивого результата. Модель может корректно генерировать текст, демонстрации выглядят убедительно, но в реальной работе ответы оказываются нестабильными, противоречивыми и не связанными с внутренними стандартами компании. Основная причина — отсутствие единого слоя знаний. В проекте для ресторанной группы с 10+ заведениями и историей более 15 лет мы сознательно начали не с агентов и не с интерфейсов, а с построения корпоративной RAG-инфраструктуры. Этот слой стал основой всей последующей AI-архитектуры.
https://habr.com/ru/articles/1039986/
#rag #искусственный_интеллект #llm #retrieval_augmented_generation #векторная_база_данных #embeddings #корпоративные_данные #ai_automation #knowledge_base #hallucinations
-
Почему RAG — фундамент любой AI-трансформации
За последние годы большинство AI-проектов в компаниях стартуют одинаково: сначала делают чат-бота, затем добавляют агентов, автоматизируют отдельные процессы и ожидают роста эффективности. На практике такие проекты часто не дают устойчивого результата. Модель может корректно генерировать текст, демонстрации выглядят убедительно, но в реальной работе ответы оказываются нестабильными, противоречивыми и не связанными с внутренними стандартами компании. Основная причина — отсутствие единого слоя знаний. В проекте для ресторанной группы с 10+ заведениями и историей более 15 лет мы сознательно начали не с агентов и не с интерфейсов, а с построения корпоративной RAG-инфраструктуры. Этот слой стал основой всей последующей AI-архитектуры.
https://habr.com/ru/articles/1039986/
#rag #искусственный_интеллект #llm #retrieval_augmented_generation #векторная_база_данных #embeddings #корпоративные_данные #ai_automation #knowledge_base #hallucinations
-
Почему RAG — фундамент любой AI-трансформации
За последние годы большинство AI-проектов в компаниях стартуют одинаково: сначала делают чат-бота, затем добавляют агентов, автоматизируют отдельные процессы и ожидают роста эффективности. На практике такие проекты часто не дают устойчивого результата. Модель может корректно генерировать текст, демонстрации выглядят убедительно, но в реальной работе ответы оказываются нестабильными, противоречивыми и не связанными с внутренними стандартами компании. Основная причина — отсутствие единого слоя знаний. В проекте для ресторанной группы с 10+ заведениями и историей более 15 лет мы сознательно начали не с агентов и не с интерфейсов, а с построения корпоративной RAG-инфраструктуры. Этот слой стал основой всей последующей AI-архитектуры.
https://habr.com/ru/articles/1039986/
#rag #искусственный_интеллект #llm #retrieval_augmented_generation #векторная_база_данных #embeddings #корпоративные_данные #ai_automation #knowledge_base #hallucinations
-
Почему RAG — фундамент любой AI-трансформации
За последние годы большинство AI-проектов в компаниях стартуют одинаково: сначала делают чат-бота, затем добавляют агентов, автоматизируют отдельные процессы и ожидают роста эффективности. На практике такие проекты часто не дают устойчивого результата. Модель может корректно генерировать текст, демонстрации выглядят убедительно, но в реальной работе ответы оказываются нестабильными, противоречивыми и не связанными с внутренними стандартами компании. Основная причина — отсутствие единого слоя знаний. В проекте для ресторанной группы с 10+ заведениями и историей более 15 лет мы сознательно начали не с агентов и не с интерфейсов, а с построения корпоративной RAG-инфраструктуры. Этот слой стал основой всей последующей AI-архитектуры.
https://habr.com/ru/articles/1039986/
#rag #искусственный_интеллект #llm #retrieval_augmented_generation #векторная_база_данных #embeddings #корпоративные_данные #ai_automation #knowledge_base #hallucinations
-
"The physical-sciences repository arXiv is banning researchers from posting their manuscripts on the platform for one year if a submission is found to contain references that have been hallucinated by artificial-intelligence tools. The ban also applies to authors who submit manuscripts containing other “incontrovertible” signs of generative AI usage that demonstrate the AI results haven’t been carefully checked.
What’s more, after a researcher’s one-year penalty is over, they will not be able to post any manuscripts to arXiv unless the work has already been accepted at a “reputable peer-reviewed venue”, according to Thomas Dietterich, a computer scientist at Oregon State University in Corvallis and chair of arXiv’s computer science section.
ArXiv’s new policy, which has triggered a torrent of both positive and negative comments from researchers on social media, is one of the latest and most far-reaching examples of how preprint servers are grappling with the rising tide of AI ‘slop’ — low-quality or meaningless content made using generative AI. Some, such as arXiv, are imposing bans on authors who do not follow their guidelines. Others have ruled out entire categories of submissions that raise concerns about generative AI use."
-
I receive a lot of email from recruiter/brokers that are looking for temporary staff (I am a software developer that works in #consultancy) and I wanted to parse the emails to make some stats on it.
Got the parsing for 80% done/readable.
Thought I could try #AI for this (never done something like it). Got less than 80% and some weird #hallucinations or interpretations. Plus the entire program takes way more time to go though all the email.
So for now my conclusion is not to use (local?) AI.
-
I receive a lot of email from recruiter/brokers that are looking for temporary staff (I am a software developer that works in #consultancy) and I wanted to parse the emails to make some stats on it.
Got the parsing for 80% done/readable.
Thought I could try #AI for this (never done something like it). Got less than 80% and some weird #hallucinations or interpretations. Plus the entire program takes way more time to go though all the email.
So for now my conclusion is not to use (local?) AI.
-
I receive a lot of email from recruiter/brokers that are looking for temporary staff (I am a software developer that works in #consultancy) and I wanted to parse the emails to make some stats on it.
Got the parsing for 80% done/readable.
Thought I could try #AI for this (never done something like it). Got less than 80% and some weird #hallucinations or interpretations. Plus the entire program takes way more time to go though all the email.
So for now my conclusion is not to use (local?) AI.
-
I receive a lot of email from recruiter/brokers that are looking for temporary staff (I am a software developer that works in #consultancy) and I wanted to parse the emails to make some stats on it.
Got the parsing for 80% done/readable.
Thought I could try #AI for this (never done something like it). Got less than 80% and some weird #hallucinations or interpretations. Plus the entire program takes way more time to go though all the email.
So for now my conclusion is not to use (local?) AI.
-
I receive a lot of email from recruiter/brokers that are looking for temporary staff (I am a software developer that works in #consultancy) and I wanted to parse the emails to make some stats on it.
Got the parsing for 80% done/readable.
Thought I could try #AI for this (never done something like it). Got less than 80% and some weird #hallucinations or interpretations. Plus the entire program takes way more time to go though all the email.
So for now my conclusion is not to use (local?) AI.
-
Youth purposely overdosing on Benadryl: Nipissing FN
Updated: May 18, 2026 at 1:32PM EDT Published: May 18, 2026 at 12:13PM EDT Nipissing First Nation’s leadership says some…
#NewsBeep #News #Medication #911 #AU #Australia #Benadryl #coma #Consciousness #death #firstnation #hallucinations #Health #high #highdoses #Nipissing #northbay #ontario #overdose #seizures #teen #Teens #TikTok #trend #youngpeople #youth
https://www.newsbeep.com/au/680315/ -
Youth purposely overdosing on Benadryl: Nipissing FN
Updated: May 18, 2026 at 1:32PM EDT Published: May 18, 2026 at 12:13PM EDT Nipissing First Nation’s leadership says some…
#NewsBeep #News #Medication #911 #AU #Australia #Benadryl #coma #Consciousness #death #firstnation #hallucinations #Health #high #highdoses #Nipissing #northbay #ontario #overdose #seizures #teen #Teens #TikTok #trend #youngpeople #youth
https://www.newsbeep.com/au/680315/ -
Youth purposely overdosing on Benadryl: Nipissing FN
Updated: May 18, 2026 at 1:32PM EDT Published: May 18, 2026 at 12:13PM EDT Nipissing First Nation’s leadership says some…
#NewsBeep #News #Medication #911 #Benadryl #coma #consciousness #Death #FirstNation #hallucinations #Health #high #highdoses #Nipissing #northbay #ontario #Overdose #seizures #Teen #Teens #TikTok #trend #UK #UnitedKingdom #YoungPeople #Youth
https://www.newsbeep.com/uk/591966/ -
https://www.europesays.com/uk/970464/ Youth purposely overdosing on Benadryl: Nipissing FN #‘high #911 #Benadryl #coma #consciousness #Death #FirstNation #hallucinations #Health #HighDoses #Medication #Nipissing #NorthBay #ontario #overdose #seizures #teen #Teens #TikTok #trend #UK #UnitedKingdom #YoungPeople #Youth
-
https://www.europesays.com/ie/491884/ Youth purposely overdosing on Benadryl: Nipissing FN #911 #Benadryl #COMA #consciousness #death #Éire #FirstNation #hallucinations #Health #high #HighDoses #IE #Ireland #Medication #Nipissing #NorthBay #ontario #Overdose #seizures #teen #Teens #TikTok #trend #YoungPeople #youth
-
Youth purposely overdosing on Benadryl: Nipissing FN
Updated: May 18, 2026 at 1:32PM EDT Published: May 18, 2026 at 12:13PM EDT Nipissing First Nation’s leadership says some…
#NewsBeep #News #Health #911 #Benadryl #CA #Canada #coma #consciousness #Death #FirstNation #hallucinations #high #highdoses #Nipissing #NorthBay #ontario #overdose #seizures #teen #Teens #tiktok #trend #youngpeople #youth
https://www.newsbeep.com/ca/678992/ -
There's a "GenAI is antisemitic because it wants me to miss the High Holidays" joke in there somewhere...
GenAI doesn't know anything: it has no concept of a fact, or a calendar, or a religion, or a tradition. All it has is associations, and it has no way to 'fact-check' or correct its errors (which is doesn't actually understand) if those errors are supported by the dataset it has absorbed.
-
There's a "GenAI is antisemitic because it wants me to miss the High Holidays" joke in there somewhere...
GenAI doesn't know anything: it has no concept of a fact, or a calendar, or a religion, or a tradition. All it has is associations, and it has no way to 'fact-check' or correct its errors (which is doesn't actually understand) if those errors are supported by the dataset it has absorbed.
-
There's a "GenAI is antisemitic because it wants me to miss the High Holidays" joke in there somewhere...
GenAI doesn't know anything: it has no concept of a fact, or a calendar, or a religion, or a tradition. All it has is associations, and it has no way to 'fact-check' or correct its errors (which is doesn't actually understand) if those errors are supported by the dataset it has absorbed.
-
🤪 Halupedia: Encyclopedia that hallucinates articles on the fly https://github.com/BaderBC/halupedia
-
LLM hallucinations in the wild: Large-scale evidence from non-existent citations
Zhenyue Zhao, Yihe Wang, Toby Stuart, Mathijs De Vaan, Paul Ginsparg, Yian Yin
"Large language models (LLMs) are known to generate plausible but false information across a wide range of contexts, yet the real-world magnitude and consequences of this hallucination problem remain poorly understood. Here we leverage a uniquely verifiable object - scientific citations - to audit 111 million references across 2.5 million papers in arXiv, bioRxiv, SSRN, and PubMed Central. We find a sharp rise in non-existent references following widespread LLM adoption, with a conservative estimate of 146,932 hallucinated citations in 2025 alone. These errors are diffusely embedded across many papers but especially pronounced in fields with rapid AI uptake, in manuscripts with linguistic signatures of AI-assisted writing, and among small and early-career author teams. At the same time, hallucinated references disproportionately assign credit to already prominent and male scholars, suggesting that LLM-generated errors may reinforce existing inequities in scientific recognition. Preprint moderation and journal publication processes capture only a fraction of these errors, suggesting that the spread of hallucinated content has outpaced existing safeguards. Together, these findings demonstrate that LLM hallucinations are infiltrating knowledge production at scale, threatening both the reliability and equity of future scientific discovery as human and AI systems draw on the existing literature."
https://arxiv.org/abs/2605.07723
#AI #GenerativeAI #LLMs #Hallucinations #AcademicPublishing #Science
-
"A 44-page report titled “Points of Attack: Uncovering Cyber Threats and Fraud in Loyalty Systems” looks like many others that management consultancy EY publishes every year.
These types of reports are the bread and butter of the biggest consultancies on the planet — thought pieces about business that help cement their position as organizations to be commissioned to help companies understand their place in the world, and how to maintain it. EY’s consulting arm generated over $16.4 billion in revenue last year.
There’s just one problem. Dig into this particular report, published in December 2025 — particularly its “resources” section, which provides links to sources of the data and claims cited throughout the document — and things don’t add up.
“It’s riddled with hallucinations,” said Edward Tian, a former employee at journalism website Bellingcat and Princeton University researcher who set up GPTZero, an AI-detection firm that shared its investigation exclusively with Sherwood News.
EY didn’t respond to multiple requests for comment.
The document is a standard advisory report describing the state of cybersecurity weaknesses in the travel industry’s loyalty points ecosystem. But it appears to have issues with citing nonexistent sources."
https://sherwood.news/tech/ai-hallucinations-appear-to-be-creeping-into-consulting-reports/
-
"A 44-page report titled “Points of Attack: Uncovering Cyber Threats and Fraud in Loyalty Systems” looks like many others that management consultancy EY publishes every year.
These types of reports are the bread and butter of the biggest consultancies on the planet — thought pieces about business that help cement their position as organizations to be commissioned to help companies understand their place in the world, and how to maintain it. EY’s consulting arm generated over $16.4 billion in revenue last year.
There’s just one problem. Dig into this particular report, published in December 2025 — particularly its “resources” section, which provides links to sources of the data and claims cited throughout the document — and things don’t add up.
“It’s riddled with hallucinations,” said Edward Tian, a former employee at journalism website Bellingcat and Princeton University researcher who set up GPTZero, an AI-detection firm that shared its investigation exclusively with Sherwood News.
EY didn’t respond to multiple requests for comment.
The document is a standard advisory report describing the state of cybersecurity weaknesses in the travel industry’s loyalty points ecosystem. But it appears to have issues with citing nonexistent sources."
https://sherwood.news/tech/ai-hallucinations-appear-to-be-creeping-into-consulting-reports/
-
"A 44-page report titled “Points of Attack: Uncovering Cyber Threats and Fraud in Loyalty Systems” looks like many others that management consultancy EY publishes every year.
These types of reports are the bread and butter of the biggest consultancies on the planet — thought pieces about business that help cement their position as organizations to be commissioned to help companies understand their place in the world, and how to maintain it. EY’s consulting arm generated over $16.4 billion in revenue last year.
There’s just one problem. Dig into this particular report, published in December 2025 — particularly its “resources” section, which provides links to sources of the data and claims cited throughout the document — and things don’t add up.
“It’s riddled with hallucinations,” said Edward Tian, a former employee at journalism website Bellingcat and Princeton University researcher who set up GPTZero, an AI-detection firm that shared its investigation exclusively with Sherwood News.
EY didn’t respond to multiple requests for comment.
The document is a standard advisory report describing the state of cybersecurity weaknesses in the travel industry’s loyalty points ecosystem. But it appears to have issues with citing nonexistent sources."
https://sherwood.news/tech/ai-hallucinations-appear-to-be-creeping-into-consulting-reports/
-
"A 44-page report titled “Points of Attack: Uncovering Cyber Threats and Fraud in Loyalty Systems” looks like many others that management consultancy EY publishes every year.
These types of reports are the bread and butter of the biggest consultancies on the planet — thought pieces about business that help cement their position as organizations to be commissioned to help companies understand their place in the world, and how to maintain it. EY’s consulting arm generated over $16.4 billion in revenue last year.
There’s just one problem. Dig into this particular report, published in December 2025 — particularly its “resources” section, which provides links to sources of the data and claims cited throughout the document — and things don’t add up.
“It’s riddled with hallucinations,” said Edward Tian, a former employee at journalism website Bellingcat and Princeton University researcher who set up GPTZero, an AI-detection firm that shared its investigation exclusively with Sherwood News.
EY didn’t respond to multiple requests for comment.
The document is a standard advisory report describing the state of cybersecurity weaknesses in the travel industry’s loyalty points ecosystem. But it appears to have issues with citing nonexistent sources."
https://sherwood.news/tech/ai-hallucinations-appear-to-be-creeping-into-consulting-reports/
-
"A 44-page report titled “Points of Attack: Uncovering Cyber Threats and Fraud in Loyalty Systems” looks like many others that management consultancy EY publishes every year.
These types of reports are the bread and butter of the biggest consultancies on the planet — thought pieces about business that help cement their position as organizations to be commissioned to help companies understand their place in the world, and how to maintain it. EY’s consulting arm generated over $16.4 billion in revenue last year.
There’s just one problem. Dig into this particular report, published in December 2025 — particularly its “resources” section, which provides links to sources of the data and claims cited throughout the document — and things don’t add up.
“It’s riddled with hallucinations,” said Edward Tian, a former employee at journalism website Bellingcat and Princeton University researcher who set up GPTZero, an AI-detection firm that shared its investigation exclusively with Sherwood News.
EY didn’t respond to multiple requests for comment.
The document is a standard advisory report describing the state of cybersecurity weaknesses in the travel industry’s loyalty points ecosystem. But it appears to have issues with citing nonexistent sources."
https://sherwood.news/tech/ai-hallucinations-appear-to-be-creeping-into-consulting-reports/
-
🚫🗞️ Oh, look! A groundbreaking #policy to #ban "hallucinated references" on #arXiv for a year. Because apparently, the ghost of #JavaScript past is plotting to take over with its band of merry unsupported browsers. 🙄🔗
https://twitter.com/tdietterich/status/2055000956144935055 #hallucinations #technews #HackerNews #ngated -
🚫🗞️ Oh, look! A groundbreaking #policy to #ban "hallucinated references" on #arXiv for a year. Because apparently, the ghost of #JavaScript past is plotting to take over with its band of merry unsupported browsers. 🙄🔗
https://twitter.com/tdietterich/status/2055000956144935055 #hallucinations #technews #HackerNews #ngated -
🚫🗞️ Oh, look! A groundbreaking #policy to #ban "hallucinated references" on #arXiv for a year. Because apparently, the ghost of #JavaScript past is plotting to take over with its band of merry unsupported browsers. 🙄🔗
https://twitter.com/tdietterich/status/2055000956144935055 #hallucinations #technews #HackerNews #ngated -
🚫🗞️ Oh, look! A groundbreaking #policy to #ban "hallucinated references" on #arXiv for a year. Because apparently, the ghost of #JavaScript past is plotting to take over with its band of merry unsupported browsers. 🙄🔗
https://twitter.com/tdietterich/status/2055000956144935055 #hallucinations #technews #HackerNews #ngated -
🚫🗞️ Oh, look! A groundbreaking #policy to #ban "hallucinated references" on #arXiv for a year. Because apparently, the ghost of #JavaScript past is plotting to take over with its band of merry unsupported browsers. 🙄🔗
https://twitter.com/tdietterich/status/2055000956144935055 #hallucinations #technews #HackerNews #ngated -
South Africa Used AI To Write Its Now Withdrawn AI Policy. The Citations Were Fake.
-
South Africa Used AI To Write Its Now Withdrawn AI Policy. The Citations Were Fake.
-
South Africa Used AI To Write Its Now Withdrawn AI Policy. The Citations Were Fake.
-
South Africa Used AI To Write Its Now Withdrawn AI Policy. The Citations Were Fake.
-
South Africa Used AI To Write Its Now Withdrawn AI Policy. The Citations Were Fake.
-
Update. "We find a sharp rise in non-existent references following widespread LLM adoption… These errors are…especially pronounced in fields with rapid AI uptake, in manuscripts with linguistic signatures of AI-assisted writing, and among small and early-career author teams. At the same time, hallucinated references disproportionately assign credit to already prominent and male scholars."
https://arxiv.org/abs/2605.07723 -
Update. "We find a sharp rise in non-existent references following widespread LLM adoption… These errors are…especially pronounced in fields with rapid AI uptake, in manuscripts with linguistic signatures of AI-assisted writing, and among small and early-career author teams. At the same time, hallucinated references disproportionately assign credit to already prominent and male scholars."
https://arxiv.org/abs/2605.07723 -
Update. "We find a sharp rise in non-existent references following widespread LLM adoption… These errors are…especially pronounced in fields with rapid AI uptake, in manuscripts with linguistic signatures of AI-assisted writing, and among small and early-career author teams. At the same time, hallucinated references disproportionately assign credit to already prominent and male scholars."
https://arxiv.org/abs/2605.07723 -
Update. "We find a sharp rise in non-existent references following widespread LLM adoption… These errors are…especially pronounced in fields with rapid AI uptake, in manuscripts with linguistic signatures of AI-assisted writing, and among small and early-career author teams. At the same time, hallucinated references disproportionately assign credit to already prominent and male scholars."
https://arxiv.org/abs/2605.07723 -
Update. "We find a sharp rise in non-existent references following widespread LLM adoption… These errors are…especially pronounced in fields with rapid AI uptake, in manuscripts with linguistic signatures of AI-assisted writing, and among small and early-career author teams. At the same time, hallucinated references disproportionately assign credit to already prominent and male scholars."
https://arxiv.org/abs/2605.07723 -
On our blog Staff SDE Yogesh Jain tries to eliminate #hallucinations in LLMs with #Postgres and the Model Context Protocol (MCP). Spoiler: he succeeds at it too! https://www.enterprisedb.com/blog/building-real-time-data-aware-intelligence-postgres-and-model-context-protocol
-
#SouthAfrica withdraws #AI #policy after it was found written by AI
source: the-independent.com/tech/ai-po…
The issue came to light when South Africa’s News24 found that at least six of the document’s 67 academic citations did not exist, while the journals they referenced were real.
Editors of the journals, including the South African Journal of Philosophy, AI & #Society, and the Journal of Ethics and Social #Philosophy, confirmed independently that the cited articles were #fake.
#politics #africa #fail #problem #technology #future #hallucinations #problem #ethics #government #news #science
-
#SouthAfrica withdraws #AI #policy after it was found written by AI
source: the-independent.com/tech/ai-po…
The issue came to light when South Africa’s News24 found that at least six of the document’s 67 academic citations did not exist, while the journals they referenced were real.
Editors of the journals, including the South African Journal of Philosophy, AI & #Society, and the Journal of Ethics and Social #Philosophy, confirmed independently that the cited articles were #fake.
#politics #africa #fail #problem #technology #future #hallucinations #problem #ethics #government #news #science
