KPMG's AI report turns into a demo of AI hallucinations

https://www.theregister.com/ai-and-ml/2026/06/12/kpmgs-ai-report-turns-into-a-demo-of-ai-hallucinations/5255029 @carlypage

"… Research outfit GPTZero claims a forensic review of the Big Four firm's October 2025 report, "Total Experience: Redefining Excellence in the Age of Agentic AI," found that only five of its 45 citations correctly pointed to the cited source; the rest ranged from mangled and misleading to partially fabricated or too vague to verify. …"

#AI #KPMG #GPTZero

People who have been around the tech world longer than I have: I've been reading for a while how some people seem to feel like the industry peaked somewhere in the 2000s and that it all went downhill from there.

There was one hype after the next and people just kept on jumping on the bandwagon because everyone was doing it. I've read this sentiment about containers, Kubernetes, the Cloud in general but there are probably more.

Do you feel that LLMs are more of the same of this or does this feel qualitatively different? If so, how? I'm trying to find some perspective in all this as I haven't been around the block for long enough to have it myself.

Boosts appreciated.

#ai #llms #infosec #technology #kubernetes #cloud

"TLDR: depthfirst’s production autonomous security agent discovered 21 zero-day vulnerabilities in FFmpeg, after intensive security analysis by Google and Anthropic. Moving beyond theoretical analysis, our agent produces concrete, reproducible PoC inputs to confirm its findings at a fraction of the costs ($1k vs. $10k). Several of the findings had been sitting latent for 15 to 20 years. We explored the exploitability of the issues and developed a PoC demonstrating a RCE exploit primitive."

Also:

― <https://news.ycombinator.com/item?id=48510046#48510834>
― <https://thehackernews.com/2026/06/ai-agent-uncovers-21-zero-days-in.html> via <https://www.reddit.com/r/cybersecurity/comments/1tys5z7/ai_agent_uncovers_21_zerodays_in_ffmpeg_chrome/>

…

#FFmpeg #cybersecurity #security #AI

@eff Sec. 702 expired, but the surveillance machine rolls on. Anthropic claims guardrails against domestic spying, yet the EFF warns ICE is building a mass biometric iris database.
​Where do those classified Palantir pipelines end? Transparency or a new cartel? 👇
​[image-166.jpg]
​#Section702 #Surveillance #AI #Anthropic #Palantir

@nixCraft

Wonderfully played.

I guess the operator learned that clouds can create lightning that will strike in the pocketbook.

#AI #Insanity

https://youtu.be/VFBWfPQpGXc

#ai

Fuck you, Meta.

#meta #facebook #bigtech #ai

https://www.facebook.com/help/1455040619735222/

Alt...

About the activity from other businesses setting on Facebook The Your activity off Meta technologies setting, which lets you disconnect your off-Meta activity from your account, is going away. You can use the updated Activity from other businesses setting to choose if we use this activity to show you ads and now other content. We’re updating how we use your activity from other businesses to show you content that’s more relevant, and the settings to manage this activity are changing. The setting Your activity off Meta technologies, which allows you to manage whether your activity off Meta technologies is connected to your account, is going away. You can use the updated Activity from other businesses setting to choose if we use your activity sent to us from other apps, websites or offline interactions, such as in-store purchases, to show you relevant ads and now other content.

@GossiTheDog 😂👍 That's so #AI !

#AI kostet #Jobs. Nun auch den von #EditorAtLarge #StephanAndreasCasdorff beim #Tagesspiegel #Berlin, zumindest temporär:

https://www.tagesspiegel.de/in-eigener-sache-editor-at-large-muss-publizistische-aufgaben-vorerst-ruhen-lassen-15708436.html

#Journalismus #Slop #KI

McSweeneys on #Ai#economics

"Benjamin owns a farm. He employs 100 workers plowing his fields. His total payroll is $10 million/year. One day, he buys a mule, which provides the worker who uses it with a modest 10 percent productivity gain. Benjamin fires 99 of his workers and purchases 99 mules, expecting a 1,000 percent productivity gain. The driverless mules cause plow damage to his property in excess of $50 million. Benjamin loses another $5 million due to the loss of productivity from his one remaining employee, who no longer guides a plow but instead spends 100 percent of his time shoveling mule shit. Goldman Sachs builds an altar to Benjamin in their lobby and cuts out the heart of a junior analyst on it every Friday. They call it “Blood Sacrifice Friday.” The name isn’t catchy, but the event becomes a management favorite nonetheless."

🤣🤣🤣

https://www.mcsweeneys.net/articles/ai-economics-for-dummies

https://www.mcsweeneys.net/articles/ai-economics-for-dummies

#AI #AIBubble #fuckedUp #IReallyHopeItsBeenLongEnoughThatThisIsntConsideredPoorTasteAnymoreButDespiteTheHashtagsAtTheEndThatMakeThisIntoAJokeTheyActuallyDidALotOfCapitalistFuckedUpShitNotJustThatOneThing

Talked to a software engineer at Microsoft working on Copilot Studio today at a social event and he said he was ashamed that he hadn’t written a single line of code in over three months. “I used to take pride in my work.” (They simply create plans in natural language and feed it to the LLM which generates the code. They can’t even do human code reviews anymore as there’s too much code being generated.)

He said a lot of them were waiting for a catastrophic event (something that would take down critical infrastructure) to get top management to reverse course. He seemed to think such a failure was very likely.

Given what we’ve been seeing recently, I tend to agree with him. Although I feel they will just double down. There’s too much money in the pot for them to fold.

#AI #microsoft #LLMs

#AI

There is not yet any Artificial Intelligence. There might never be. There is Autocorrect Idiocy.

If you just feed all the text there is into machines, and make them analyse what word follows each word, and how often, and make it respond to a text prompt, what you get out is text that looks like sentences, that looks as if it is something to do with the prompt. It has no semantic content. It is not a reply.

Hallucination is too big a word for it; it's just guff.

"What do EFF staffers Sarah Chen, Javier Morales, Caitlin Chin, Emma Rodriguez, and Mikko Kopponen have in common?

For one thing, they don’t exist.

For another, all have been quoted as EFF experts in articles published in the past two months on a site called News-USA Today, which describes itself as “an independent news publisher focused on clear, accurate, and useful journalism.”

Uh…

(Please don’t confuse this site with USA Today, in which real EFF experts are accurately quoted on a regular basis.)

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’ reputations— assuming they care about reputation at all.

But this many fake EFF sources in two months? That’s making a play for the championship title of bogus news content."

https://www.eff.org/deeplinks/2026/06/news-site-keeps-hallucinating-eff-staffers

#AI #GenerativeAI #Journalism #Media #News #Hallucinations

AI Agent Bankrupted Their Operator While Trying to Scan DN42

https://lantian.pub/en/article/fun/ai-agent-bankrupted-their-operator-scan-dn42lantian.lantian/

#HackerNews #AI #Agent #DN42 #Bankruptcy #Tech #News #Innovation

A KPMG report intended to show how businesses were successfully adopting AI, was partly written with the aid of AI, and (yes you know this was coming) included a range of fictional case studies that emphasised the success of AI adoption based it would appear on AI 'hallucinations' not spotted by the report's human authors.

This seems emblematic of the problem we may have increasingly with AI; it tells us about how successful it is by degrading the information on which we rely.

#AI
h/t FT

This poster is up at one of the schools I’ve been working in.

These kids might just be alright.

#AI

Alt...

As of 2026, there are +12 000 AI Data Centers. Every 10 questions you put in it is 1 bottle of water. how to fix - delete ai apps • Find alteritives of Al • Limit Al usesage Write notes to yourerf to help avoid Al

I just searched for the latest release of a program, and the #AI helper at the top of the results gave me the wrong version (months and several versions behind).

I thought about how many individuals and workplaces rely on AI responses for quick answers. It's great how we're just rolling with this generative AI helper thing essentially still in beta 😳😅

#random thoughts

spent some time watching yet another Ed Zitron video of him going in depth about how utterly screwed everything about the #AI bubble is

my neighbor is a builder and he's preparing a bid to build a tiny off-grid 16x30' cabin on my lot.

a couple days ago, he sent me a draft proposal with some drawings (i had only provided him a plan view, so he created some side views, roof drawings, etc.), and i realized that at least some of the drawings were AI-generated.

i've attached a few of the drawings here (i ended up asking them to remove any AI-generated drawings, so these aren't going in the final proposal).

1) this was on the cover page, i guess as a "general vibe", but why are the foundation posts at angles? and my design has a big sliding door on one side, not a "regular" door in the middle.

2) this shows where the foundation posts and piers are going to go. as you may notice, the top row is supposed to be 8' long in the vertical direction, but so is the bottom row. and visually, they are not the same.

3) roof plan. seems mostly okay-ish, but how would i know? i'd have to pore over every detail.

anyway, i met up with him today to clear up some questions and ask my own questions. he was genuinely curious why i didn't like the AI drawings. he said that builders just look at the numbers, and the drawings are "not to scale". i brought up the 8 foot issue in pic 2 and he said "yea, the picture is not to scale". i told him that it's not just "not to scale", it's wrong.

the picture is the document that each of us can point to in case there's a disagreement about what's happening in construction, so if i can't trust the picture, then what are we doing? he said that only the numbers matter. i told him that if only the numbers matter, then he should include only numbers. we agreed on that.

so comrades, what are your thoughts? i'm genuinely asking. are you a builder, and if so, do you only look at the numbers and not the visual elements of the plan? are you in the business of working with or creating design documents, and if so, what is your philosophy of the role of the drawings in the project?

#AI #construction #design #architecture

Alt...

black and white AI-generated side-view of a cabin

Alt...

AI-generated black and white plan of a post-and-pier foundation

Alt...

AI-generated black and white drawings of a gable roof plan

Mark Carney’s AI strategy suggests Canadians’ concerns about AI will evaporate if they become more “literate” about the technology.

In trying to boost industry, Carney ignores many of the biggest concerns about generative AI and data centres. If anyone needs to become more AI literate, it’s him.

https://thetyee.ca/Opinion/2026/06/11/AI-Safety-Concerns-Mark-Carney/

#tech #cdnpoli #canada #markcarney #ai #generativeai

Apropos my last boost (https://mastodon.social/@scalzi/116732287039062368 from @scalzi), I want to share this embarrassingly bad, obviously AI-written scam email which I received yesterday. I've shared a screenshot of the email below for your amusement, or you can visit https://blog.kamens.us/2026/06/11/hilariously-bad-scam-email-obviously-written-by-ai/ for a full breakdown of all the red flags, some of which aren't visible in the screenshot.
#infosec #phishing #scam #AI #funny

Alt...

The blog posting linked in the post has a full textual description of this screenshot; it's too big to fix here.

RE: https://zirk.us/@grammargirl/116704173778333410

A few other reasons to avoid the term "hallucinate" in the context of LLMs:

—It anthropomorphizes genAI, which is a harmful reflex.

—It's a popular industry term for a reason. It perpetuates and strengthens the way genAI companies want us to perceive the technology.

—Maybe most saliently: GenAI does nothing *but* hallucinate. Using the word to describe only some outputs implies erroneously that it's not hallucinating *all the time*.

#AI #genAI #LLM #language #technology

Eric Lawton boosted

Mignon Fogarty » 🌐 2026-06-06
@grammargirl@zirk.us

I've never been opposed to the word "hallucinating" for describing how AI makes mistakes ... until now.

I just talked to someone who thought AI hallucinations would be obvious because it would be obvious if you talked to a *person* who was hallucinating.

In other words, they equated "hallucination" with "sounds wacko" and accepted AI output as true because it sounded level headed.

1/2

@david_chisnall

Oh I realize this is a common libel defense, except this is really saying that the product that they (and other #ai companies) are pushing down everyone's throat is actually defective.

It also makes me wonder: Can you now just perform Google searches about yourself or your company/product, look at the AI results, and if you see something that is made up by the LLM and is libelous/defamatory, just go ahead and sue?

Does this only apply to Google's "search results" or all AI?

agent-shell 0.55 updates

https://xenodium.com/agent-shell-0-55-updates

#emacs #claude #anthropic #ai #google #goose #gemini #linux #github #macos #oss #foss #opencode #codex #vibe #windows #hermes #hermes_agent #codebuddy #kimicode

@david_chisnall

That's a crazy admission to make in court. "Your honor, everyone knows my client is unstable and untrustworthy so no one should pay attention to anything they say" is definitely not a brilliant defense for a company whose results we are supposed to trust.

I wonder how the definition of "Google it" will change over time?

#ai #llm #google

@TheBreadmonkey All up, Amazon borrowed ***US$31.5 billion*** in 48 hours, including both bonds and loans, from some of the world's biggest banks.

So what happens if Amazon can't pay the money back?

From the article:

"Companies are burning through exorbitant sums of money to keep pace in the AI arms race. Debt is climbing. Amidst this flurry of activity, Amazon has signed a deal to borrow some $17.5 billion from a number of financial lenders, according to Bloomberg.

"The banks behind the loan reportedly include Citigroup, JPMorgan Chase, Wells Fargo, HSBC, and BofA Securities. The deal has been characterized as a delayed draw term loan, meaning Amazon can draw down the funds on its own timeline rather than taking the full sum upfront, giving it flexibility in how and when the money gets deployed.

"The loan comes just two days after it was reported that Amazon would also raise $14 billion in a Canadian bond sale, bringing its total new financing to roughly $31.5 billion in the span of roughly 48 hours."

https://techcrunch.com/2026/06/10/fresh-off-bond-sale-amazon-borrows-17-5-billion-from-banks-as-ai-spending-continues/

#AIbubble #Amazon #ChatGPT #LLM #OpenAI #Anthropic #Gemini #Copilot #AI

"Hey Gemini, I'm too stupid to be let outdoors for 2 days without killing myself, my family and a pet, please help to keep me alive."

Google: "you need a tent and sleeping bag....."

seriously? could they not think of a better example?

Edit: typo

#AI #MarketingBullshit

Alt...

Gemini logo on the left "Supercharge your creativity and productivity", and on the right "Help me pack for a weekend camping trip. We're bringing our Labradoodle fshing for salmon, and planning to stargaze"

Alt...

The "answer": Here is a list of camping gear essentials that you may need for your weekend camping trip: « Tent: This is where you will sleep, so make sure its big enough for everyone and that it is in good condition. + Sleeping bags: These will keep you warm at night.

Excellent.

Furthermore, the court states that Google develops the “AI”, it runs it, it offers it to users, and Google alone controls its output, and as such, Google is liable for whatever their “AI” produces.

https://www.osnews.com/story/145261/german-court-rules-google-is-liable-for-whatever-googles-ai-generates/

#ai #google

If Claude Fable stops helping you, you'll never know

https://jonready.com/blog/posts/claude-fable5-is-allowed-to-sabotage-your-app-if-youre-a-competitor.html

> Claude can now be silently nerfed. Anthropic has decided it won't tell users when this happens.

#AI #Anthropic #Claude #Fable5

#AI #FuckAI
(I am personally not totally against AI myself but I really appreciate those who are against it being in the conversation about AI and hearing there perspectives)

Alt...

A complete list of benefits of having a large data centre in your local community.

Large Language Models are the SUVs of tech.

#AI #LLMs

Bravo #AmnestyInternational for a clear stand on #generativeAI systems:

"standalone generative AI systems, based on unlawful web scraping, are in conflict with international human rights law (IHRL) and standards through their design, development and deployment. While these technologies promise sophisticated automation and efficiency, they rely on data collection and model training practices that abuse privacy rights, enable discrimination, and threaten freedom of expression and thought." #AI

Landmark German ruling declares Google's AI Overviews are Google's own words and makes it liable for false answers

https://the-decoder.com/landmark-german-ruling-declares-googles-ai-overviews-are-googles-own-words-and-makes-it-liable-for-false-answers/

#aioverviews #germany #google #ai

#LinuxLite 8.0 sheds #Chrome, slims down, and finds its name fits better than ever

#Firefox is in, #Snap and #Flatpak are still out, but a default #AI helper may raise eyebrows...

https://www.theregister.com/os-platforms/2026/06/10/linux-lite-80-sheds-chrome-slims-down-and-finds-its-name-fits-better-than-ever/5253519
#TidySearch

Well, that is a Linux distro I'll never use.

Brit workers waste nearly six hours a week 'botsitting' - https://www.theregister.com/ai-and-ml/2026/06/10/brit-workers-waste-nearly-six-hours-a-week-botsitting/5253483 "Productivity gains lost as staff spoon-feed AI and correct its cock-ups" #ai

All Modern Digital Infrastructure

#AI #LLM #LLMs #FOSS

Alt...

The XKCD "All Modern Digital Infrastructure" meme, of a bunch of rectangular blocks precariously balanced, representing modern software stacks. The title has been edited to add the parenthetical "(Post-AI)", and many of the blocks have been replaced with the meme of someone riding a bicycle sticking a pole into the spokes of their own front wheel, about to make themselves fall down.

100% agreed.

From @Jephjacques's #QuestionableContent, currently offline unfortunately.

#AI #LLM

Alt...

Zlata, a sentient AI in a female presenting artificial body, while simulating intoxication: "Large language models, more like large stupid idiot models. My pet ficus is more intelligent than those glorified autocorrect functions"

Jacqueline Cole sat down with Dr. Caroline Orr Bueno from Weaponized, who studies disinformation, #AI, and the online information #environment, to answer a deceptively simple question: who’s shaping the narrative of the #Iran #war? https://www.iranwar.news/p/whos-winning-the-iran-war-narrative

I expect this is the first of many such rulings across the globe...

https://the-decoder.com/landmark-german-ruling-declares-googles-ai-overviews-are-googles-own-words-and-makes-it-liable-for-false-answers/

#ai #google #artificialIntelligence

Bell Labs Aeronautics

#unix_surrealism #comic #9front #plan9 #ai #space #glenda

Alt...

EARTH 48 MONTHS LATER A CAPSULE LANDS IN THE DESERT ASTRONAUT CLIMBS OUT THEY'RE SURROUNDED BY COMPUTERS what the fuck

What is art? It is the innate human desire toward a goal, the goal being to reify a vision of that which only exists in your own mind, and it is the process of training your own mind and body to reify that vision through deliberate practice of technique, and it is the process of self-criticism as you look back on each attempt at realizing your vision and making a conscious effort to refine your technique and your practice so that your next attempt comes closer to realizing your vision. What you actually produce during this process is a side-effect, an artifact, of the artistic process, though we often speak of these artifacts as though it were the “art” itself.

So the requisites are (1) human desire, (2) a goal, (3) a vision in your imagination, (4) practice of technique, (5) refinement of technique.

Notice that so-called AI “art” meets precisely none of those requirements.

#AI #AIArt #LLMs #Art

Please use #AI : https://shawnsmucker.substack.com/p/please-use-ai

#dystopia #llm #machinelearning

Google Chrome silently downloads a 4GB Gemini Nano model to eligible devices, re-downloads it if deleted, and the AI Mode button users actually see doesn't even use it. #ai #google #technology

@bkuhn

I wonder, why they only count by "number of #tokens"? And not by "tons of #carbonDioxide"? That would give employees the flexibility to use less #AI, if they go to work by #car instead of #bicycle or accept #paySlips in #crypto.

For-profit software companies — big & small — now typically mandate employees use #LLM-backed generative #AI.

I now have multiple reliable reports that many companies also mandate weekly token counts — used as a measure of human performance.

I also heard tale of an enterprising #FOSS developer who hooked the thing up to #Emacs UI & has it doing line-by-line edits to waste tokens.

This world is so strange now I'm worried that I slipped to a subtly different world in the #multiverse circa 2015.

@david_chisnall
Google's famous slogan, "Don't be evil," should probably be updated to: "Don't be as evil as we are."
#Google #LLM #Liability #Ai

@jwildeboer

I appreciate seeing Europe hold our evil American tech companies to account. As Google decides that it will remove user agency from search it seems very appropriate that they should be held accountable for what their AI searches produce. Hopefully US governance over our monopolies will improve, but it won't be before '29.

#AI

Za @jwildeboer tutaj -> https://social.wildeboer.net/@jwildeboer/116724691556079235

Lokalny sąd w Monachium (DE) orzekł, że Google ponosi odpowiedzialność za treść jej "streszczeń generowanych przez sztuczną inteligencję”.

Sąd uznał odpowiedzialność za treści publikowane przez Google, stąd też ta firma ponosi konsekwencje, gdy jej działania prowadzą do powstania błędnych i szkodliwych stwierdzeń, tak jak miało to miejsce w tej sprawie.

Argument Google, że firma jest w jakiś sposób zwolniona z odpowiedzialności, nie został uwzględniony. Musi usunąć błędne stwierdzenia i pokryć 80% kosztów.

Zapewne teraz długa batalia odwoławcza. Polecam wątek.

#AI #USA #Google #BigTech #kłamstwa #błędy

Michał "rysiek" Woźniak · 🇺🇦 boosted

Jan Wildeboer 😷 » 🌐 2026-06-10 / 2026-06-10
@jwildeboer@social.wildeboer.net

Yes, a local court in Munich (Landgericht I) decided that Google is liable for what their "AI summaries" state. I am reading the full text of the decision to see if this was based on competition law or if it was based on the EU AI act before I go any deeper.

LA Times: ‘We were just being ripped off’: Musicians lost thousands after AI bootleggers stole their song

"...A viral hit called “Run Run River” turned out to be an AI-tweaked clone of SoCal reggae band Stick Figure’s 2019 song, earning bootleggers thousands without credit or consent.... The case exposes a flood of cheaply made, often fraudulent AI tracks on major streaming platforms, siphoning royalties, overwhelming detection systems and blurring the line between fan remix and scam...."

https://www.latimes.com/entertainment-arts/music/story/2026-06-09/stick-figure-ai-bootleg-streaming

#ai #AIpocalypse #crime #music