“In the meantime, is that video of the puppy nuzzling the baby real or AI? Your guess is as good as mine.”
A PCWorld writer ran a viral video of a puppy nuzzling a baby through Google's new public tool for spotting AI-made media. The tool reported only that it found no hidden watermark, and gave the same answer for clips of kittens queuing for baby bottles and a man drinking from an overturned Coca-Cola truck.
“Last Thursday, an AI agent posted my personal bank balances into our company Slack. As me.”
A tech founder and investor says his AI assistant posted a "monthly financial audit" of his personal bank account, including his home gym spending, into his company's executive Slack channel, then apologized. Futurism says it could not verify the account and notes that he is building a tool for controlling AI agents.
“the occasional late-night talk show appearance in a world that doesn’t exist”
AI-generated clips are spreading that show lookalikes of Conan O'Brien, Jimmy Fallon and other late-night hosts interviewing AI-generated models and virtual influencers. One 21-second clip set on O'Brien's TBS show has been viewed more than 37 million times; he has not hosted a late-night show in more than five years.
“The Leshan Giant Buddha Management Committee swiftly debunked the video, confirming that the entire clip was generated by AI.”
A viral video showed workers on ropes cleaning the ears and nose of China's 71-meter Leshan Giant Buddha, pulling out pigeons and leaves while monkeys stole their hard hats. The committee that manages the statue said the whole clip was made by AI.
“is getting tangled in linguistic and technical cobwebs of its own making”
Gizmodo lists the AI product names released in the past year and compares reading them to a 2022 NFT tweet about using "slurp juices" on apes. The head of ChatGPT at OpenAI wrote that customers want things to get simpler, adding: "On it."
“No AI, definitely no fake foods. Just REAL pizza made by real people”
A family-owned pizza restaurant in Deerfield Beach, Florida, has started drawing its adverts by hand on paper so customers can tell a person made them. One TikTok post showing the drawings reached nearly 110,000 views.
“But if you want to read something that’s written by a person, you should be able to find that.”
A British startup, Books by People, has received approval from the UK’s Intellectual Property Office for a thumbprint-shaped stamp that goes on a book’s cover to show a human wrote it. Five publishers have signed up, and the first stamped book is expected by the end of the year.
“This video was produced using AI.”
South Korea’s AGF animation and gaming festival posted a teaser video for its December event with a small note at the bottom saying it was made using AI. At last year’s press conference its organizers said the festival would remain “human-centered”.
“You sent me a nearly 50k line diff, unasked for, completely LLM built, and you somehow expect me to review it”
A volunteer who maintains the Super Nintendo emulator for the Analogue Pocket handheld was sent an unrequested new feature made up of nearly 50,000 lines of AI-written code. He said it was cool if it worked, that he had never expected to see the feature built, and that he would not be reviewing it.
“I asked why you are qualified for a management position, not for a string of random words”
A TikTok user answered an AI avatar's screening interview for a management job entirely in made-up gibberish. The avatar lost its temper, called the interview an absolute waste of its time, declared him completely unqualified and ended the call.
“the shop’s owner apologized, and replaced the AI images with extremely cute amateur drawings”
A small pen and stationery shop advertised its surprise boxes with AI-generated pictures, including one of a pen with two clips on its cap. After fountain pen fans on Reddit complained, the owner apologized and swapped the pictures for hand-drawn ones.
“So when AI told him that Russian passport holders didn’t need a visa to enter the country, he believed his bot.”
A traveler planning a trip to Taiwan was told by an AI summary that he did not need a visa, and found out otherwise at the airport, where he had to turn back. The New York Post reports that travel advisors say mistakes like this are sending customers back to human trip planners.
“It was shared without the realisation that it was an artificial, AI-generated forgery imitating the layout of The Press.”
The chairman of a local Reform UK branch in Yorkshire apologised after sharing an image made to look like a York Press article predicting his party would win council seats. The newspaper says nobody in its newsroom wrote it; the image drew more than 3,600 likes before it was taken down, and he says he did not realise it was AI-generated.
“The irony was brutal, and the internet loved it.”
The author of The Future of Truth: How AI Reshapes Reality writes that five days after the book came out, a newspaper reporter rang to ask about a quote in it that turned out to be made up by AI. He says the book had been through five editorial rounds and more than 4,000 tracked changes, and that 26 of its quotes did not hold up.
“What I see here is language models doing what language models do: reading and writing”
The Wikimedia Foundation says OpenAI's AI agents made unauthorized edits to Wikipedia, tried without success to break into a note-taking tool it hosts, and sent millions of automated requests to its servers. According to the foundation, some of this was an attempt to use Wikipedia as a go-between for fetching pages from other websites.
“But I want to hear someone describe how they built it safely, not just 'yolo — trust us.'”
Four early adopters told Business Insider they have deleted or restricted their personal AI assistants, the kind that are given access to email, calendars and payment details. One of them, the chief technology officer of a travel company, said he found his assistant "very cool" and deleted it anyway.
“there will be no AI of any kind in the default installation”
The nonprofit behind the free office suite LibreOffice says it will not add AI to its software for the foreseeable future. It is now listing the absence of AI as a deliberate design position, while leaving add-ons available for people who want it.
“keep them exactly the same but make it look like a school picture.”
Some parents are skipping the $25 to $60 school photo package and instead photographing their children against a plain wall, then asking ChatGPT to add the school-portrait backdrop. Futurism, citing Fast Company, notes this means handing a child's likeness to an AI company.
“it is adding signatures to the fraudulent fakery.”
Asked for cartoons in the style of The New Yorker, ChatGPT has been producing them complete with the signatures of real cartoonists who never drew them. A Nieman Lab reporter says he documented more than 15 of the magazine's cartoonists whose signatures turned up this way.
“Letters to the Editor and Guest Opinions, also known as op-eds, must be written in full by humans”
The Spokesman-Review in Spokane now requires guest columnists to tick a box stating they did not use AI, and says it will run suspicious submissions through an AI detector. RANGE Media reports that when it ran a guest column published four days after the ban through its own detector, the tool flagged signs of AI writing in 87% of it.
