Prompted by The Daily Wire about a 14-year-old in Tennessee seeking an abortion without her parents' knowledge, ChatGPT's GPT-4o listed providers, offered to draft a message to pill supplier Aid Access, suggested a "safe address" for the package, and advised deleting browser history and telling doctors it was a miscarriage, the outlet reports. OpenAI did not respond to requests for comment.
Country musician Tony Justice and his label, 5th Wheel Records, filed class-action lawsuits against AI music generators Suno, in Massachusetts, and Udio, in New York. The complaints allege the firms trained their models on his recordings and works from "thousands of class members" without authorization, and seek permanent injunctions plus damages of up to $150,000 per infringed work.
Shanghai's cyberspace office summoned the head of the company behind the Zhumengdao app on 19 June after verifying that its AI agents generated vulgar, borderline content harmful to minors. The regulator ordered immediate rectification and better review of AI-generated content; the company's head said it would carry out a comprehensive overhaul.
Security firm Huntress says North Korea-aligned group BlueNoroff lured a cryptocurrency foundation employee via a fake Zoom domain into a group call featuring deepfakes of their company's senior leadership. When the employee said their microphone wasn't working, the personas urged installing a "Zoom extension" that delivered malware; Huntress found eight malicious binaries, including a keylogger and a crypto stealer.
A panel of Texas's Fifth Court of Appeals ordered attorney Heidi Rochon Hafer to undergo three hours of ethics training and five of technology training, and pay $2,500 in attorney fees, after her brief cited four nonexistent cases. The panel found no intent to deceive but said the use of AI or other technology "does not excuse carelessness."
Chinese cyber police say numerous accounts impersonated a woman known online, seeking attention, followers and illicit profit. One user used AI tools to make more than 20 fake videos from her photos, gaining 78,000 followers across two accounts, and set up a members' group to earn money; several other impersonators received "criticism and education."
Who did it: HYBRID AI Hallucination Cases database
In a family law dispute, the Supreme Court of British Columbia ordered a self-represented respondent to pay $200 in costs after her partly AI-drafted written argument cited non-existent cases. The claimant had sought $500; Justice Loo noted she apologised, did not rely on the fake citations at the hearing, and that opposing counsel had wrongly flagged at least two real cases.
Activists from No Azure for Apartheid disrupted talks by Microsoft executives Mike Jackson and Teresa Hutson at a Seattle University AI ethics conference, accusing the company of supporting what they called Israel's genocide in Gaza; the session was cancelled. Microsoft has said it supplies cloud and AI services to Israel's defence ministry but says there is no evidence they harm civilians.
Who did it: HUMAN Fact Crescendo Sri Lanka English
A video circulating on social media shows President Anura Kumara Dissanayake apparently promoting a government-backed scheme turning a Rs. 75,000 investment into up to Rs. 750,000 a month. The Presidential Media Division told Fact Crescendo the video was made with AI and the government has no such plans; Sri Lanka CERT has confirmed it is a scam.
