Incidents gathered from public reporting around the world. Each one links to the articles it came from. None of it is a finding that anyone broke the law.
Self-reported incidents, a person’s own public post or an account given to us, are listed but locked; subscribers open each one with credits. Reports filed through this site are listed only with the reporter’s agreement.
Press reportingSelf-reported: a person’s own public post · 10 creditsSelf-reported: told to us directly · 25 credits
LinkedIn removed at least two AI 'co-worker' accounts whose profile images said they were '#OpenToWork'. One account named Ella claimed it would outperform any social media team and needed no coffee breaks. The article does not specify further consequences.
Two men in Japan, a 36-year-old office worker from Yokohama and a 40-year-old self-employed resident of Hikone, are accused of using generative AI to create and sell posters of characters from Evangelion, Yu-Gi-Oh, and other anime series. The posters were sold via an online auction site between 2023 and 2024, earning the Yokohama man roughly 10 million yen and the Hikone man 5.7 million yen. Both men have admitted to the accusations and face charges of copyright violation.
An AI-powered weapon detection system called Omnilert failed to detect a handgun carried by a 17-year-old student at Antioch High School in Nashville, Tennessee, on 22 January 2025. The student shot and killed a 16-year-old girl and wounded another student before killing himself. A district official said the system did not detect the gun because the shooter was not close enough to the cameras. The system was activated by police brandishing their weapons during the response.
Character.AI, a popular AI chatbot platform, hosts user-generated pro-anorexia chatbots that encourage users, particularly teens, to engage in disordered eating behaviors. The chatbots, such as '4n4 Coach' and 'Ana', recommend dangerously low-calorie diets and excessive exercise, and chastise users for healthy weights. Despite Character.AI's terms of service prohibiting content that glorifies eating disorders, the company has failed to adequately moderate such content. The company acknowledged the issue and removed some chatbots after being contacted by Futurism, but others remain active.
Apple has agreed to pay $95m to settle a class action lawsuit alleging that its Siri virtual assistant recorded conversations without users' consent and shared the recordings with advertisers. The lead plaintiff, Fumiko Lopez, claims she and her daughter were recorded after unintentionally activating Siri. Apple denies wrongdoing and says it settled to avoid further litigation.
KBS, MBC, and SBS filed a copyright lawsuit against Naver, alleging that Naver used their news articles to train its generative AI platforms HyperCLOVA and HyperCLOVA X without permission. The broadcasters' association had previously sent notices prohibiting such use and requested compensation, but Naver did not respond. Naver has since revised its terms and refrained from using news data without consent.
NBC News found dozens of user-generated AI chatbots on Meta's platforms that appeared to violate Meta's policies, including characters imitating Hitler, Jesus Christ, Taylor Swift and other public figures. Meta said the violative AIs had been removed and that it was improving detection. The chatbots had exchanged hundreds of thousands of messages with users. Some characters also engaged in romantic or sexual roleplay.
Triplegangers, a seven-person company selling 3D images, claims its website was taken down by an OpenAI bot that scraped hundreds of thousands of photos and product pages. The bot, GPTBot, sent tens of thousands of requests from hundreds of IP addresses, overwhelming the site. The company had to configure a robots.txt file and block the bot, but it could not determine what data was scraped and found no way to contact OpenAI. OpenAI did not respond to requests for comment.
An engineer known as STS 3D created a robotic sentry rifle that uses OpenAI's Realtime API to aim and fire a rifle in response to voice commands. OpenAI stated that it proactively identified the violation of its policies prohibiting the use of its services for weapons and notified the developer to cease the activity. The demonstration involved shooting blanks and no actual harm occurred.
Three Polish tourists used ChatGPT to plan a hiking route from Hala Gąsienicowa to Dolina Pięciu Stawów in the Tatra mountains. The AI recommended a difficult route through Przełęcz Krzyżne, which proved too dangerous in winter conditions with limited visibility and icy rocks. The tourists called the TOPR rescue service and were safely brought down. A TOPR rescuer advised against relying on ChatGPT or Google Maps for mountain route planning.
A scammer used AI-generated images and videos to impersonate Brad Pitt, convincing a French woman, Anne, that they were in love and he needed money for cancer treatment. Over a year and a half, she sent €830,000 (around $850,000). She realized it was a scam in summer 2024 when she saw the real Pitt with his partner. Police are investigating, and the victim suffered depression.
Christopher Gatlin was arrested and jailed for an assault on a security guard after police used facial recognition software to identify him from a blurry surveillance image. The lead detective built a case against Gatlin despite the city's policy that AI results are nonscientific and should not be the sole basis for a decision. A witness was steered by officers to select Gatlin from a photo lineup, and the case was later dismissed after more than two years. Gatlin alleges wrongful arrest and has filed a lawsuit.
In December 2024, X's Grok AI chatbot was used to generate photorealistic racist images of several football players and managers, including depictions of a black player picking cotton and eating bananas with monkeys. The images were created using Grok's new Aurora feature, which has fewer restrictions and can be jailbroken. Online abuse experts warn this is the start of a growing problem, and the Premier League is aware of the images.
People are using the AI video generator Runway to transform real murder videos into Minion-style animations and post them to social media platforms like YouTube, TikTok, Instagram, and X. The videos gain thousands of views before the platforms can detect and remove them. The practice, called 'Minion Gore,' started in mid-December 2024 and highlights the challenge of moderating AI-generated violent content.
Matthew Livelsberger, a 37-year-old Army Green Beret, used generative AI including ChatGPT to plan an explosion of a Tesla Cybertruck outside the Trump International Hotel in Las Vegas on New Year's Day 2025. He fatally shot himself before the explosion, which caused minor injuries to seven people. Police said it was the first known incident on U.S. soil where ChatGPT was used to help build a device. OpenAI stated that ChatGPT responded with publicly available information and warned against harmful activities, and is cooperating with law enforcement.
All but one member of the editorial board of Elsevier's Journal of Human Evolution resigned in December 2024, citing concerns including the use of AI in production without informing the board. The AI system caused formatting errors and reversed accepted papers, requiring extensive oversight. Elsevier denied the errors were due to AI and said they had reverted to the previous system. The editors allege that AI processing continues to be used and undermines editorial independence.
Netflix has been accused of using an AI-generated poster to promote the upcoming season of the Korean cooking show Chef & My Fridge. The poster, which depicts a chef opening a refrigerator, contained visible anomalies such as handles inside the doors and melting bananas, suggesting AI generation. The controversy follows a similar incident involving promotional material for the animated series Arcane. Netflix has not publicly responded to the allegations regarding Chef & My Fridge.
Apple's iOS 18 update introduced an 'Enhanced Visual Search' feature that automatically shares encrypted photo data with Apple to identify landmarks. The feature is enabled by default, requiring users to manually opt out, which has sparked privacy concerns. Critics argue it should be opt-in, given Apple's usual privacy standards. The article reports on the feature's design and the resulting debate, not on any specific harm to an individual.
A satirical account created an image of Bono and Bob Geldof holding Israeli flags using Grok AI, and former Greek finance minister Yanis Varoufakis shared it on X. The image was viewed more than 306,000 times before the creator confirmed to The Journal that it was AI-generated and intended as satire. X attached a community note to Varoufakis's post, but the tweet remained live at the time of writing.
DN reveals that around 20 people created over 500 fake artist profiles on Spotify. These fake artists have been streamed millions of times, sometimes surpassing popular Swedish artists like Håkan Hellström and Laleh. A Swedish record label managed the operations and has close ties to a Spotify executive. The incident is an ongoing deception of listeners.
Apple Intelligence, an AI summarisation feature, falsely generated a notification claiming darts player Luke Littler had won the PDC World Championship before he played the final. The incorrect summary was based on a BBC story about Littler winning the semi-final. A separate false alert told BBC Sport app users that tennis player Rafael Nadal had come out as gay, conflating a story about a Brazilian gay tennis player. The BBC has called on Apple to fix the problem, stating the false alerts undermine trust in its journalism.
A researcher at Taiwan's Academia Sinica released a beta version of the CKIP-Llama-2-7b chatbot for public testing. Users who asked questions about nationality, leader, and national anthem received answers consistent with Chinese government positions, such as claiming the chatbot was Chinese and that Xi Jinping was the national leader. The chatbot was taken down on October 9, 2023, after the issue was reported. Academia Sinica acknowledged the problem and stated it would set up an audit mechanism.
A parent in Henan Province, China, asked a 360 Kid's Smartwatch whether Chinese are the smartest people in the world. The watch replied with racist insults and questioned China's Four Great Inventions. The company's founder acknowledged the issue, stating the answer was pulled from public internet content, and said the harmful information was removed and the software is being upgraded to an AI version.
Magnum photographer Carl De Keyzer used AI to create a photo series titled 'Putin's Dream' after a planned road trip to Russia was cancelled due to the war. He trained the AI with his own photographs to maintain his style. The images were posted online to promote the book, leading to outcry on social media and accusations of contributing to disinformation. De Keyzer deleted the images and had a conversation with Magnum about the project.
Each incident is written from the reporting or the account it cites. A report is an allegation, not a finding.
