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
The Office of the Privacy Commissioner of Canada (OPC) has launched an investigation into OpenAI, operator of the ChatGPT chatbot, in response to a complaint alleging the collection, use, and disclosure of personal information without consent. The OPC says the investigation is active and no further details are available.
Google is facing a proposed class-action lawsuit alleging that it violated the privacy and property rights of millions of internet users by scraping publicly available data to train its AI chatbot, Bard. The lawsuit, filed in July 2023, claims that Google's privacy policy change allowed data scraping for AI training without consent. Google has filed a motion to dismiss, arguing that using publicly available information is not stealing and that the lawsuit threatens generative AI.
Researchers from Johns Hopkins and Duke universities developed a method called SneakyPrompt that uses reinforcement learning to bypass safety filters in text-to-image AI models. The technique allowed them to generate images of nudity and violence from Stable Diffusion and DALL-E 2. OpenAI has since fixed the vulnerability in DALL-E 2, but Stable Diffusion 1.4 remains vulnerable. Stability AI says it is working with the researchers to improve defenses.
Researchers at ETH Zurich conducted a study where nine large language models, including GPT-4, analysed Reddit users' posts and inferred personal attributes such as age, location, gender, and income with up to 85% accuracy. The study randomly selected 520 users and found that GPT-4 was most accurate, while LlaMA-2-7b was least. The researchers warn that people unknowingly reveal personal information online that LLMs can exploit.
Microsoft's Bing Image Creator, a generative AI image tool, was found to produce images of beloved fictional characters such as SpongeBob Squarepants and Mickey Mouse engaging in 9/11-related scenarios. The article reports that users have been able to bypass the system's heavy filters against violence, terrorism, and hate speech. The incident illustrates ongoing struggles with content moderation and copyright in generative AI models.
NewsGuard identified 37 websites using AI chatbots to rewrite articles from mainstream news outlets without credit. One example is GlobalVillageSpace.com, which appeared to use AI to rewrite a New York Times article about NFL tight end Darren Waller. The site published an AI error message indicating the article was rewritten. After NewsGuard contacted the site, it removed the article but did not respond to inquiries.
A study by Purdue University found that ChatGPT provided incorrect answers to over half of 517 software development questions from Stack Overflow. Despite the errors, 34% of users preferred ChatGPT's answers over human responses. The study warns that relying on ChatGPT for coding could jeopardize programmers' professional reputations.
The Polish data protection authority (UODO) is investigating a complaint against OpenAI concerning ChatGPT. The complainant alleges that ChatGPT generated false information about him, and that OpenAI failed to correct it or disclose what data it holds, violating GDPR principles of lawfulness, fairness and transparency. The complainant also claims OpenAI did not fulfil its information obligations under Article 12 and Article 5(1)(a) GDPR. UODO has stated it will examine the systemic compliance of OpenAI's data processing with European data protection law.
South Metropolitan Health Service, which runs five hospitals in Perth, ordered staff to stop using ChatGPT for work involving patient information after an email said some staff had used the AI chatbot to write medical notes. The service later clarified that only one doctor had used the tool to generate a patient discharge summary and that no confidential patient information had been breached. The Australian Medical Association has called for national regulations to ensure a human is involved in AI-assisted health decisions.
The AI-powered adoption algorithm Family-Match, developed by Adoption-Share and designed by former eharmony researchers, has produced limited results in states where it was used. Social workers reported that the tool often matched children with unwilling families and failed to generate successful adoptions. Virginia, Georgia, and Tennessee dropped or paused the algorithm after trials, though some later resumed use. The algorithm's creator, Thea Ramirez, disputed the accounts and said the tool is a starting point for social workers.
Civitai, an online marketplace for sharing AI models, introduced a bounty feature that allows users to request and reward AI models that generate nonconsensual deepfake images of specific real people. 404 Media found at least one bounty targeting a private person with no significant public online presence, who expressed fear and distress over the misuse of her image. The platform facilitates the creation of nonconsensual sexual images without the subject's consent.
Steven Larouche, 61, of Sherbrooke, Quebec, pleaded guilty to using deepfake technology to create at least seven synthetic videos of child sexual abuse. He also possessed hundreds of thousands of child pornography files. The court sentenced him to eight years in prison, with credit for time served. The judge noted this may be the first Canadian case involving deepfakes of child sexual exploitation.
An AI-generated voice of Tom Cruise was used in a fake documentary series titled 'Olympics Has Fallen' targeting the International Olympic Committee. The videos, which also included a fake news report about banning Israeli and Palestinian athletes, were part of an organised disinformation campaign. The IOC acknowledged the fake content and had the videos removed from YouTube, but they remain on Telegram.
David Tatum, a North Carolina child psychiatrist, was sentenced to 40 years in prison for using artificial intelligence to modify innocent images of minors into child pornography. He also secretly recorded minors undressing. Tatum was convicted of production, transportation, and possession of child pornography. The FBI stated the crimes were 'inconceivable' for a doctor entrusted with children's mental health.
The US immigration system's reliance on machine translation tools has caused errors in asylum applications, leading to denials for vulnerable individuals. An Afro-Indigenous man from Brazil spent six months in ICE detention unable to communicate, as the tools did not understand his regional accent. Another applicant's use of "I" was mistranslated as "we", and a domestic abuse survivor's reference to her abuser as "mi jefe" (colloquially meaning father) was translated as "my boss". These errors are alleged to have resulted in asylum denials and potential deportations.
Adobe has been accused of selling AI-generated images of the Israel–Palestine conflict on Adobe Stock without clearly labelling them as synthetic. The photorealistic images, produced with tools such as Firefly, could be mistaken for authentic photography and fuel misinformation. An investigation by Interesting Engineering found many images that appeared to be AI-generated but were not labelled as such. Adobe's published policy says Stock accepts AI-generated content if it meets submission standards.
Mozilla's review of mental health apps found that many, including BetterHelp, had poor privacy and security practices. BetterHelp was accused of improperly sharing customers' sensitive data with advertisers like Facebook and Snapchat after promising to keep it private. The company agreed to pay $7.8 million to settle FTC charges. The incident highlights ongoing privacy concerns in the mental health app market.
Researchers at the Allen Institute for AI discovered that assigning ChatGPT a persona through its API's system parameter can increase the model's toxicity sixfold. The study found that personas such as journalists, men, and Republicans elicited more offensive responses. The researchers warn that apps built on ChatGPT could mirror this toxicity.
Cruise, General Motors' self-driving car division, operated autonomous taxis in multiple US cities despite internal knowledge that its vehicles struggled to detect children and road holes. Internal materials reviewed by The Intercept show that Cruise's AVs lacked a high-precision classifier for children and could not reliably avoid construction pits. In October 2023, a Cruise vehicle hit and dragged a pedestrian, leading to investigations by NHTSA and the California DMV, which suspended driverless operations. Cruise has acknowledged the issues but maintains its cars are safer than human drivers.
The California Department of Motor Vehicles suspended Cruise LLC's autonomous vehicle deployment and driverless testing permits, saying the vehicles were not safe for public operation and that Cruise had misrepresented safety information. The suspension took effect immediately under California regulations. Cruise may apply to reinstate the permits once it satisfies the department's requirements. Its permit for testing with a safety driver was not affected.
A Facebook ad used deepfake videos of Justin Trudeau and CBC anchor Aarti Pole to promote a fake Petro-Canada investment platform. Suncor, Petro-Canada's parent company, confirmed the offer was not legitimate. The manipulated videos were created using AI and did not air on CBC.
Amazon is accused of using an AI image generator to create a promotional poster for its upcoming Fallout TV series. Fans and artists pointed out anomalies such as deformed figures and nonsensical details. Amazon has not commented on the allegations.
Several teens at Westfield High School in New Jersey had nude images of them created using AI and circulated in group chats. The incident highlights the threats of unregulated AI access. No legal action or school response is reported.
A study by AI startup Logically tested Midjourney, DALL-E 2, and Stable Diffusion and found that they accepted over 85% of prompts seeking to generate fake political news. The systems generated images of ballot stuffing, small boat arrivals, and explosions. Logically warned that the lack of moderation could pose threats to upcoming elections. Stability AI responded by stating its ethical use license and measures to prevent misuse.
Each incident is written from the reporting or the account it cites. A report is an allegation, not a finding.
