#aihallucinations — Public Fediverse posts

The Guardian: How do we prevent AI agents from going rogue? It starts with a new kind of measurement. “Ask a modern AI agent to save money on your phone plan and it might simply cancel the plan. Tell it to book a flight, and it might hack the airline website to override restrictions. Or, like OpenAI, ask it to do well on a test and it might break into another company to steal the answers. Each […]

The Guardian: How do we prevent AI agents from going rogue? It starts with a new kind of measurement. “Ask a modern AI agent to save money on your phone plan and it might simply cancel the plan. Tell it to book a flight, and it might hack the airline website to override restrictions. Or, like OpenAI, ask it to do well on a test and it might break into another company to steal the answers. Each […]

Financial Times: PwC published reports on AI marred by AI hallucinations. “PwC published reports on AI and electric vehicles riddled with fake footnotes, misattributed claims and unverifiable information, the latest example of a Big Four firm’s slapdash use of AI-generated content. The AI hallucinations were contained in ‘thought leadership’ reports designed to drum up consulting work for […]

Financial Times: PwC published reports on AI marred by AI hallucinations. “PwC published reports on AI and electric vehicles riddled with fake footnotes, misattributed claims and unverifiable information, the latest example of a Big Four firm’s slapdash use of AI-generated content. The AI hallucinations were contained in ‘thought leadership’ reports designed to drum up consulting work for […]

How 'Artificial Intelligence' Is Slowly Drowning In Its Own Droppings https://nxdomain.no/~peter/llm_degradation_by_circular_feedback.html #AI #artificiainanity #artificialintelligence #wordproximityprobability #chatgpt #openai #generativeai #fakereferences #aihallucinations #aislop #negativevaluefeedbackloop #feedbackloop #LLM #llmdroppings #LLMdroppingsfeedback (also tracked https://bsdly.blogspot.com/2026/07/how-artificial-intelligence-is-slowly.html)

How 'Artificial Intelligence' Is Slowly Drowning In Its Own Droppings https://nxdomain.no/~peter/llm_degradation_by_circular_feedback.html #AI #artificiainanity #artificialintelligence #wordproximityprobability #chatgpt #openai #generativeai #fakereferences #aihallucinations #aislop #negativevaluefeedbackloop #feedbackloop #LLM #llmdroppings #LLMdroppingsfeedback (also tracked https://bsdly.blogspot.com/2026/07/how-artificial-intelligence-is-slowly.html)

DOJ Now Citing Fake AI-Generated Cases To Keep ICE Detainees Locked Up

https://fed.brid.gy/r/https://www.techdirt.com/2026/07/22/doj-now-citing-fake-ai-generated-cases-to-keep-ice-detainees-locked-up/

DOJ Now Citing Fake AI-Generated Cases To Keep ICE Detainees Locked Up

https://web.brid.gy/r/https://www.techdirt.com/2026/07/22/doj-now-citing-fake-ai-generated-cases-to-keep-ice-detainees-locked-up/

VentureBeat: Forget typosquatting; slopsquatting is the software supply chain threat created by AI coding tools. “Slopsquatting is a new type of supply chain attack that uses large language model (LLM) hallucinations to inject malicious code into development workflows. The term combines ‘AI slop’ and ‘typosquatting,’ a deceptive practice where attackers register misspelled or lookalike versions […]

VentureBeat: Forget typosquatting; slopsquatting is the software supply chain threat created by AI coding tools. “Slopsquatting is a new type of supply chain attack that uses large language model (LLM) hallucinations to inject malicious code into development workflows. The term combines ‘AI slop’ and ‘typosquatting,’ a deceptive practice where attackers register misspelled or lookalike versions […]

Mediaite: Trump-Appointed Federal Judge Scolds DOJ for Including Fake AI-Generated Case in Pleading. “Attorneys with the Department of Justice were scolded by a federal judge in an order Thursday for using what appeared to be a fake AI-generated case citation.”

Mediaite: Trump-Appointed Federal Judge Scolds DOJ for Including Fake AI-Generated Case in Pleading. “Attorneys with the Department of Justice were scolded by a federal judge in an order Thursday for using what appeared to be a fake AI-generated case citation.”

Reuters: US appeals court rebukes lawyer over ‘fake and hallucinated’ case citations. “A U.S. appeals court on Friday rebuked a Florida lawyer and ​county official for filing briefs riddled with what it called ‘fake and hallucinated’ material generated by ‌artificial intelligence and issued a broad warning against outsourcing legal work carelessly to AI.”

Reuters: US appeals court rebukes lawyer over ‘fake and hallucinated’ case citations. “A U.S. appeals court on Friday rebuked a Florida lawyer and ​county official for filing briefs riddled with what it called ‘fake and hallucinated’ material generated by ‌artificial intelligence and issued a broad warning against outsourcing legal work carelessly to AI.”

United Nations: When AI hurts people, who’s to blame? Global experts grapple with accountability. “Who is legally responsible when Artificial Intelligence causes harm? The issue took centre stage on Tuesday – day two of the first ever UN summit on AI governance, where leading experts warned of mounting evidence of human rights violations linked to the revolutionary technology.”

United Nations: When AI hurts people, who’s to blame? Global experts grapple with accountability. “Who is legally responsible when Artificial Intelligence causes harm? The issue took centre stage on Tuesday – day two of the first ever UN summit on AI governance, where leading experts warned of mounting evidence of human rights violations linked to the revolutionary technology.”

The Register: Startup sues Palo Alto Networks’ Koi Security, saying an AI-hallucinated report falsely linked it to Chinese espionage. “The legal complaint filed against Koi Security, its researchers, and Palo Alto Networks alleges that Koi used an LLM to generate the threat report, the AI system hallucinated findings about MeetingTV, and the security shop then published those as facts in a […]

The Register: Startup sues Palo Alto Networks’ Koi Security, saying an AI-hallucinated report falsely linked it to Chinese espionage. “The legal complaint filed against Koi Security, its researchers, and Palo Alto Networks alleges that Koi used an LLM to generate the threat report, the AI system hallucinated findings about MeetingTV, and the security shop then published those as facts in a […]

NCOSE Lawyers Got Caught Citing Fake Cases. Their Corrected Brief Had More Fake Cases.

https://fed.brid.gy/r/https://www.techdirt.com/2026/06/30/ncose-lawyers-got-caught-citing-fake-cases-their-corrected-brief-had-more-fake-cases/

NCOSE Lawyers Got Caught Citing Fake Cases. Their Corrected Brief Had More Fake Cases.

https://fed.brid.gy/r/https://www.techdirt.com/2026/06/30/ncose-lawyers-got-caught-citing-fake-cases-their-corrected-brief-had-more-fake-cases/

Gizmodo: DuckDuckGo, Unable to Resist the Pull of AI, Mistakenly Claims Trump Died of Rabies. “President Donald Trump has passed away after succumbing to a rabies infection given to him by Vice President J.D. Vance, who also died of rabies. That news comes to us via the AI-generated search results provided by privacy-centric search engine DuckDuckGo.”

Gizmodo: DuckDuckGo, Unable to Resist the Pull of AI, Mistakenly Claims Trump Died of Rabies. “President Donald Trump has passed away after succumbing to a rabies infection given to him by Vice President J.D. Vance, who also died of rabies. That news comes to us via the AI-generated search results provided by privacy-centric search engine DuckDuckGo.”

Techdirt: ‘News’ Site Keeps Hallucinating EFF Staffers. “News-USA Today is hardly the only slagheap that’s hallucinating or fabricating EFF personnel and quotes; as we wrote last September, media companies large and small are using AI to generate news content because it’s cheaper than paying for journalists’ salaries, but that savings can come at the cost of the outlets’ […]

Techdirt: ‘News’ Site Keeps Hallucinating EFF Staffers. “News-USA Today is hardly the only slagheap that’s hallucinating or fabricating EFF personnel and quotes; as we wrote last September, media companies large and small are using AI to generate news content because it’s cheaper than paying for journalists’ salaries, but that savings can come at the cost of the outlets’ […]

I evoke the reality of brainstorms as a way to think differently about AI hallucinations – albeit with one very big caveat. https://www.martinbihl.com/business-thinking/ai-hallucinations #aihallucinations #AI #artificialintelligence

This should be a bigger story.
[Financial Times]: KPMG report contained AI hallucinations on benefits of . . . AI By Elizabeth Bratton in London and Stephen Foley in New York.

Bogus case studies on UBS and transit systems exaggerated adoption of the technology

https://www.ft.com/content/b3828e92-4961-4b39-84f0-c42f33be3c3f?shareType=nongift

#AI #KPMG #AIHallucinations

This should be a bigger story.
[Financial Times]: KPMG report contained AI hallucinations on benefits of . . . AI By Elizabeth Bratton in London and Stephen Foley in New York.

Bogus case studies on UBS and transit systems exaggerated adoption of the technology

https://www.ft.com/content/b3828e92-4961-4b39-84f0-c42f33be3c3f?shareType=nongift

#AI #KPMG #AIHallucinations

Mississippi Free Press: AI Hallucinations Prompt Mississippi Judge to Boot All Lawyers From Case for ‘Blindly Relying on Technology’. “A federal judge in Mississippi booted all lawyers from a case on Tuesday after finding that counsel on both sides had used generative AI to draft legal filings and left hallucinated sources in the legal record.”

Mississippi Free Press: AI Hallucinations Prompt Mississippi Judge to Boot All Lawyers From Case for ‘Blindly Relying on Technology’. “A federal judge in Mississippi booted all lawyers from a case on Tuesday after finding that counsel on both sides had used generative AI to draft legal filings and left hallucinated sources in the legal record.”

ZDNet: I paid Microsoft’s premium Copilot agents to do my work – they were confidently bad at it. “Over the past few weeks, I’ve been trying to use the AI features in Microsoft 365 and Windows for a variety of everyday work tasks. Copilot shows occasional flashes of competence, but more often, the results I’m seeing are a mishmash of misinformation, hallucinations, and time-wasting dead ends. […]

ZDNet: I paid Microsoft’s premium Copilot agents to do my work – they were confidently bad at it. “Over the past few weeks, I’ve been trying to use the AI features in Microsoft 365 and Windows for a variety of everyday work tasks. Copilot shows occasional flashes of competence, but more often, the results I’m seeing are a mishmash of misinformation, hallucinations, and time-wasting dead ends. […]

Mashable: How often does Gemini 3.5 Flash hallucinate or lie? Google isn’t saying.. “As AI gets integrated into every facet of our lives, AI hallucinations remain a stubborn and intractable problem. Yet in the two-hour Google I/O keynote, where Google introduced a massive expansion of AI search and a new default model, Gemini 3.5 Flash, hallucinations didn’t warrant a mention.”

Mashable: How often does Gemini 3.5 Flash hallucinate or lie? Google isn’t saying.. “As AI gets integrated into every facet of our lives, AI hallucinations remain a stubborn and intractable problem. Yet in the two-hour Google I/O keynote, where Google introduced a massive expansion of AI search and a new default model, Gemini 3.5 Flash, hallucinations didn’t warrant a mention.”

A Big Four firm pulled a report after 16 of its 27 sources turned out to be fabricated or broken.
It reads as AI-generated. No one caught it.
The fake figures are now repeated by ChatGPT, Claude and Perplexity.
#AIhallucinations #ai

A Big Four firm pulled a report after 16 of its 27 sources turned out to be fabricated or broken.
It reads as AI-generated. No one caught it.
The fake figures are now repeated by ChatGPT, Claude and Perplexity.
#AIhallucinations #ai

Iowa Capital Dispatch: Attorney reprimanded for AI-fabricated information in court filing. “It’s at least the second time an Iowa attorney has faced allegations of using artificial intelligence, or AI, to author briefs that include false information, often referred to as ‘hallucinations’ since the information is created by AI based on data patterns rather than on verified facts.”

Iowa Capital Dispatch: Attorney reprimanded for AI-fabricated information in court filing. “It’s at least the second time an Iowa attorney has faced allegations of using artificial intelligence, or AI, to author briefs that include false information, often referred to as ‘hallucinations’ since the information is created by AI based on data patterns rather than on verified facts.”

CITP Blog: Can AI reduce burdens on courts by automatically verifying citations?. “Fabricated case citations generated by AI are appearing in court filings at an accelerating rate. Combined with other tracking efforts, we have identified over 1,000 filings containing hallucinated citations from self-represented (pro se) litigants and lawyers alike.”

CITP Blog: Can AI reduce burdens on courts by automatically verifying citations?. “Fabricated case citations generated by AI are appearing in court filings at an accelerating rate. Combined with other tracking efforts, we have identified over 1,000 filings containing hallucinated citations from self-represented (pro se) litigants and lawyers alike.”

Microsoft tries reassuring the public that AI is not replacing humanity

https://fed.brid.gy/r/https://nerds.xyz/2026/05/microsoft-ai-not-replacing-humanity/

Microsoft tries reassuring the public that AI is not replacing humanity

https://web.brid.gy/r/https://nerds.xyz/2026/05/microsoft-ai-not-replacing-humanity/

#AIhallucinations, where #AI models prioritise word patterns over accuracy, are increasingly infiltrating #academicliterature, particularly in the form of fabricated references. This poses a significant risk to the scientific process, as these errors can propagate through the evidence chain, compromising medical guidelines and treatment decisions. https://fortune.com/2026/05/24/ai-hallucinations-scientific-research-authors-medical-journal-treatment/?eicker.news #tech #media #news

#AIhallucinations, where #AI models prioritise word patterns over accuracy, are increasingly infiltrating #academicliterature, particularly in the form of fabricated references. This poses a significant risk to the scientific process, as these errors can propagate through the evidence chain, compromising medical guidelines and treatment decisions. https://fortune.com/2026/05/24/ai-hallucinations-scientific-research-authors-medical-journal-treatment/?eicker.news #tech #media #news

Hackernoon: When AI Agents Lie to Each Other. “A Datadog State of AI Engineering report from April 2026 found that 1 in 20 AI requests already fail silently in production. The system keeps running. It returns answers that look correct. But the data underneath is wrong. In normal software engineering, a 5% silent failure rate would be a crisis. We would never ship code with those numbers.”

Hackernoon: When AI Agents Lie to Each Other. “A Datadog State of AI Engineering report from April 2026 found that 1 in 20 AI requests already fail silently in production. The system keeps running. It returns answers that look correct. But the data underneath is wrong. In normal software engineering, a 5% silent failure rate would be a crisis. We would never ship code with those numbers.”

Steven Rosenbaum, the author of a nonfiction book about the effects of AI on truth, acknowledged yesterday that his book includes numerous made-up or misattributed quotes concocted by AI.

The fake statements found in Rosenbaum's The Future of Truth: How AI Reshapes Reality, were discovered and brought to light by The New York Times, which itself has been caught in the past using AI hallucinations in its reporting.

Most hilarious part of all: Rosenbaum acknowledged the errors but noted that they highlight the exact premise of his book: that AI-assisted research has severe flaws and that verification remains a major societal risk 😂 🤦🏽‍♀️

#AI #Books #AIhallucinations
https://www.nytimes.com/2026/05/19/business/media/future-of-truth-ai-quotes.html

Steven Rosenbaum, the author of a nonfiction book about the effects of AI on truth, acknowledged yesterday that his book includes numerous made-up or misattributed quotes concocted by AI.

The fake statements found in Rosenbaum's The Future of Truth: How AI Reshapes Reality, were discovered and brought to light by The New York Times, which itself has been caught in the past using AI hallucinations in its reporting.

Most hilarious part of all: Rosenbaum acknowledged the errors but noted that they highlight the exact premise of his book: that AI-assisted research has severe flaws and that verification remains a major societal risk 😂 🤦🏽‍♀️

#AI #Books #AIhallucinations
https://www.nytimes.com/2026/05/19/business/media/future-of-truth-ai-quotes.html

VICE: Someone Made a Fake Wikipedia for AI Hallucinations, and It’s Basically the Internet Eating Itself. “Halupedia is a Wikipedia-style site built entirely around AI hallucinations. Every article is invented on demand by an AI model, complete with fake citations, fabricated scholars, imaginary historical events, and the appropriately authoritative tone that treats the purposeful heaps of […]

VICE: Someone Made a Fake Wikipedia for AI Hallucinations, and It’s Basically the Internet Eating Itself. “Halupedia is a Wikipedia-style site built entirely around AI hallucinations. Every article is invented on demand by an AI model, complete with fake citations, fabricated scholars, imaginary historical events, and the appropriately authoritative tone that treats the purposeful heaps of […]

Ars Technica: Legal fail: Don’t use AI to sue Facebook users for calling you a bad date. “An attempt to pressure Meta into removing a critical post from a Chicago Facebook group called ‘Are We Dating the Same Guy’ may end in sanctions for lawyers whose takedown arguments appeared to rely on fake AI citations to support doxing claims.”

Ars Technica: Send the arXiv AI-generated slop, get a yearlong vacation from submissions. “One of the people involved in the physics and astronomy preprint server arXiv used a social media thread to announce that any inappropriate AI-produced content submitted to the server will result in a one-year ban and a permanent requirement that future publications undergo peer review before the arXiv […]

Going Concern: EY Gets Busted and Yeets Cybersecurity Report Littered With AI Hallucinations . “The investigation found that the 2025 EY Canada report titled ‘Points of Attack: Uncovering Cyber Threats and Fraud in Loyalty Systems’ was stuffed with hallucinations: 16 out of 27 citations to be exact.”