The daily rap sheet of the machine age.
12 items
Digital media publisher Ziff Davis, whose outlets include ZDNet, PCMag, CNET and IGN, sued OpenAI in Delaware federal court on April 24, accusing it of misusing its publications to train the model behind ChatGPT. The complaint says OpenAI "intentionally and relentlessly" exploited copyrighted content; OpenAI says its models are trained on publicly available data and grounded in fair use.
Original report: Ziff Davis sues OpenAI over ChatGPT training
South Korea's Personal Information Protection Commission said DeepSeek transferred user data to several firms in China and the US without users' consent or disclosure, including AI prompts and device, network and app information sent to Beijing Volcano Engine Technology Co. According to the commission, DeepSeek stopped sending prompts on April 10; the regulator recommended it immediately destroy the transferred prompt data.
Original report: Korean regulator says DeepSeek sent user data abroad without consent
The Korean Newspaper Association filed a complaint with the Fair Trade Commission on April 24, alleging Naver used news content without permission to train its HyperCLOVA and HyperCLOVA X models and refused to disclose details. The association also claims Naver's AI search copied articles, distorted their context and omitted or falsified sources, abusing its dominant market position.
Original report: Korean newspapers accuse Naver of training AI on their news
Brazil's social security institute's AI-powered app, Meu INSS, has rejected claims from hundreds of people in remote areas with little digital literacy over minor errors, Rest of World reports; one former sugarcane worker's retirement claim was turned down because the system identified her as a man. INSS says automated decisions follow legal criteria; appeals take 278 days on average.
Original report: Brazil's AI benefits app turns down claims over minor errors
A federal magistrate judge in the Eastern District of New York fined lawyer Victoria Hovsepyan $1,000 after her reply brief in a negligence suit against Costco cited five cases that do not exist. She admitted generating them with an AI platform called ChatOn and reviewing its work for "maybe twenty minutes"; the court found subjective bad faith.
Original report: Lawyer fined $1,000 after ChatOn invents five cases
Indonesia's national police cybercrime directorate handed two suspects, identified as AMA and JS, and their case files to prosecutors over deepfake videos of President Prabowo Subianto and other officials that, police say, offered government aid on condition recipients first sent money. AMA was arrested on January 16, 2025, and JS on February 4.
Original report: Indonesian police hand Prabowo deepfake scam suspects to prosecutors
Police and the county sheriff's office in Iowa are investigating after high school students allegedly used AI apps to create fake nude images of several female classmates; how many images and victims there were is unknown. The school district says it disciplined those responsible, and parents have started a petition for updated school policies.
Original report: Iowa students accused of making AI nudes of classmates
Okta Threat Intelligence says facilitators of schemes placing North Korean nationals in remote tech jobs use generative AI services to generate CVs, rehearse interviews, test applications against automated screening and manage multiple personas, helping minimally skilled workers keep roles long enough to send earnings to the sanctioned regime. Suspected personas have used real-time deepfake video in interviews.
Original report: Okta says North Korean job schemes rely on generative AI
Indonesian police arrested three people in Pangandaran, West Java, over a scam in which, police say, they ran accounts resembling East Java Governor Khofifah Indar Parawansa and used deepfake videos to offer victims cheap motorbikes. The national police cybercrime director promised more cyber patrols and firm sanctions.
Original report: Police arrest three over deepfake governor motorbike scam
Malaysia's Prime Minister's Office said investigations are under way into several inaccurate depictions of the national flag, believed to stem from AI-generated images, published by a newspaper, a foreign company exhibiting in Kuala Lumpur and the Ministry of Education's SPM performance report. Speaking for Anwar Ibrahim, his press secretary promised no double standards and said AI cannot replace human editorial judgment.
Original report: Malaysia investigates flag errors blamed on AI image generation
Asda has received more than 5,000 complaints over a live facial recognition trial at five Greater Manchester stores, following a campaign by Big Brother Watch, which called it "deeply disproportionate and chilling". The system checks shoppers' faces against a list of people who previously engaged in criminal activity at Asda sites; Asda has been contacted for comment.
Original report: Asda facial recognition trial draws more than 5,000 complaints archived copy of Asda facial recognition trial draws more than 5,000 complaints
Weibo users complained that its AI search feature, Weibo Zhisou, summarised ordinary users' online personas from their posts; on April 22 it analysed one low-follower user's health management and emotional state for Red Star Capital Bureau, but not a day later. Weibo said on April 23 it by default draws on the past month of public posts and denied infringement.
Original report: Weibo's AI search draws privacy complaints for profiling users
