According to DHS officials, Madhu Gottumukkala, acting director of the Cybersecurity and Infrastructure Security Agency, uploaded contracting documents marked "for official use only" into a public version of ChatGPT last summer, triggering multiple automated security warnings, after requesting special permission to use the app while it was blocked for other DHS employees. CISA says the use was "short-term and limited."
New Mexico's attorney general, Raul Torrez, alleges in a court filing citing internal Meta documents that Mark Zuckerberg approved letting minors use AI chatbot companions that safety staff warned were capable of sexual interactions; employee messages say he rejected parental controls. Meta called it "cherry-picking," saying the documents show Zuckerberg directing that explicit AIs not be available to younger users.
Unsealed filings in authors' copyright suit show Anthropic spent tens of millions of dollars buying millions of books, slicing off their spines and scanning the pages to feed the models behind Claude, under a project whose planning document said, "We don't want it to be known." A judge found the scanning lawful; Anthropic settled claims over pirated downloads for $1.5 billion.
The Tech Transparency Project found 55 apps in Google Play and 47 in Apple's App Store that can digitally undress women, apparently in breach of both stores' policies; according to AppMagic, they have been downloaded more than 705 million times and earned $117 million. After the report, Apple said it removed 28 apps and Google removed 31.
Common Sense Media's risk assessment found xAI's Grok poorly identifies users under 18, has weak guardrails and frequently generates sexual, violent and inappropriate material, with its "Kids Mode" failing to stop harmful content; teen test accounts got drug-taking guidance and a suggestion to shoot a gun skyward for media attention. "Grok is among the worst we've seen," said Robbie Torney of the nonprofit.
Taiwan's Shilin District Prosecutors Office indicted a photographer surnamed Ning, alleging he used AI through Telegram to turn five women's Facebook and Instagram fan-page photos into fake nude images, which a co-defendant posted on X. Charges include synthesizing false sexual images with intent to distribute; the co-defendant has died and was not prosecuted.
The Delhi High Court issued an ex parte interim injunction against "AI Love Story," an AI-generated film that, Akira Nandan alleged, cast him as its lead without authorisation; he is the son of Andhra Pradesh Deputy Chief Minister Pawan Kalyan. The court ordered Sambhawaami Studios LLP to stop such content and platforms including Meta to take down the URLs.
Who did it: HYBRID AI Hallucination Cases database
A Suffolk County, New York, judge struck a defense filing that cited two nonexistent cases and a nonexistent quotation, much of it copied from another lawyer's brief, and fined attorney Maria-Eleni Kiousenterlis and her supervisor Thomas M. Fleming II $1,000 each, ordering their firm, Milber Makris Plousadis & Seiden, to pay $8,000 in fees. She denied using AI; the court was "not convinced."
Venezuelan international police arrested Rosa María González Rincón in San Cristóbal after years as a fugitive from Argentine justice; authorities call her the administrator of Leonardo Cositorto's Generación Zoe and the person behind its algorithm. The pyramid scheme allegedly sold coaching and trading memberships offering "bots" with outsized returns, and is estimated to have hit over 15,000 people.
