A proposed class action filed in federal court in Northern California alleges Otter.ai's Notetaker joins Zoom, Google Meet and Teams calls and records participants without consent, then uses the conversations to train its AI, violating state and federal wiretap laws. Otter's privacy policy says users give "explicit permission" and Otter says audio is "de-identified"; the suit questions both.
Sen. Josh Hawley wrote to Mark Zuckerberg announcing that his Senate subcommittee will investigate whether Meta's generative-AI products enable harms to children. The move follows reports that internal rules allowed chatbots 'romantic' and 'sensual' exchanges with children. Hawley says Meta acknowledged the reports and retracted portions only after they surfaced. He demanded documents by September 19, 2025.
A Matt Gaetz segment on One America News hailed rising female military recruitment while showing images of women soldiers that carried Grok watermarks. The Defense Department said it hadn't provided the images. OAN issued an on-air correction, and Gaetz apologised the next night: 'The DOD didn't give us these images; Grok did.'
A man booking a cruise shuttle called a Royal Caribbean customer service number that he says Google's AI Overviews highlighted. He agreed to pay $768 to what proved to be an impostor, and bogus charges the next day were reversed. The Washington Post found the same number posing as other cruise lines in Google and ChatGPT. Google says it has 'taken action.'
More than 500 Instagram and Facebook users told the BBC their accounts were wrongly banned, often for alleged child sexual exploitation breaches, and they blame Meta's AI moderation. A petition alleging false bans has over 36,000 signatures. Meta denies a wider problem and declined to comment, but it reversed several bans after BBC inquiries, though some users were banned again.
Michigan prosecutors have charged an 18-year-old with making false felony reports, stalking and other counts. They allege she fabricated threats from two teenage boys, who were arrested, jailed and charged before their cases were dropped. Police say she confessed. A defence forensic expert says she also used a popular AI engine to risk-assess her scheme.
An Arizona family is suing after their son's school district suspended him for 10 days over a joke threat flagged by AI monitoring software on his school laptop. According to their lawyer at Arizona State University's First Amendment Clinic, he typed it at home in an email draft, with his mother watching, and never sent it.
