Apple was sued on Wednesday in a US federal court in a proposed class action alleging its iPhone 16 marketing promised Apple Intelligence upgrades to Siri that, according to the complaint, "did not exist then and do not exist now." The suit alleges false advertising, fraud and breach of contract. Apple has said some Siri personalization improvements are delayed until 2026.
A Norwegian man with no public profile asked ChatGPT about himself and was told he had murdered his two young sons and received a 21-year sentence. In a complaint to Norway's data protection authority, he and digital rights group Noyb allege the "defamatory" answer breaches GDPR accuracy rules and seek a fine; OpenAI says it concerns a version since upgraded with web search.
A federal judge in Illinois gave final approval to Clearview AI's settlement of lawsuits alleging it scraped billions of online photos into a facial recognition database in breach of Illinois' biometric privacy law. The class, expected to number 125,000, gets a 23% stake in the company, which the judge estimated at $51.75 million, over objections from 22 states and the District of Columbia.
Paul Roberts, former CEO of ad-fraud detection firm Kubient, is due to serve one year and one day in prison after pleading guilty to securities fraud. Prosecutors say he schemed to create $1.3 million in fraudulent revenue to bolster its IPO and had staff produce fake reports for its AI tool, KAI, "based on made-up metrics".
Who did it: HYBRID AI Hallucination Cases database
At a March 12 hearing in a lawsuit over a death in Wayne County Jail, a federal judge in Michigan admonished the plaintiff's lawyer for AI-generated bogus legal citations and ordered an ethics seminar. The judge's March 21 ruling, dismissing the case, also called the brief's misstatement of medical records a "brazen factual fabrication", "perhaps also" from AI, and warned of disciplinary action.
Police in Nanchong, Sichuan, say a 35-year-old man used AI software on March 17 to concoct a post claiming a Gaoping district middle school canteen made buns from lymph meat, poisoning 87 students. The school says nothing of the sort happened; police say he wanted attention and followers. He was administratively penalised and his account banned.
An AI-generated video of market analyst Michael Hewson, host of the Good Money Guide Podcast, circulated on Facebook and Instagram urging viewers to join a WhatsApp group for "massive undervalued" investments and to "get in before the crowd". Hewson said it was "nothing to do with me", that he is taking it up with Meta and is "very angry".
Sir Stelios Haji-Ioannou, founder of easyGroup, warns that an AI-generated video falsely using his name and image is running in Facebook and Instagram ads that point to websites unconnected to him or the easy brands. Sir Stelios called it a serious fraud designed to part the public from their savings and vowed to leave no stone unturned finding those responsible.
Who did it: HYBRID AI Hallucination Cases database
A federal judge in Kentucky said there was reason to believe a self-represented plaintiff used AI to draft his amended complaint against Hilton, which cited one case that does not exist and another that did not say what he claimed. The court warned such citations may bring sanctions while dismissing his defamation suit over an email calling a thread's tone "rather aggressive".
An audio clip in which a Trump-like voice announces renaming the "District of Columbia" to the "District of America" spread on TikTok, Facebook, Bluesky and X among critics and some supporters, one post drawing 1 million views. Digital forensics experts told NewsGuard it is AI-generated; the now-deleted TikTok account that first posted it apparently made other fake Trump clips.
Cofense says a phishing campaign sends fake Instagram ad-suspension emails from a Salesforce address, leading Meta Business users to a lookalike page where a fake support chatbot asks for screenshots and personal details. Failing that, a bogus "System Check" guide tricks victims into adding the attackers' authenticator app as a two-factor login method.
A viral video reused Business Today footage of Reliance Industries director Anant Ambani speaking at Vantara, overlaid with synthetic audio, to promote a gaming app "Avi by Ambani" via a fake Google Play page, Newschecker found. ElevenLabs' detector rated the audio 98% likely made with ElevenLabs; the Deepfakes Analysis Unit called it a bid to scam the public.
A viral AI filter that edited photos to make people look as though they had put on weight was removed from TikTok after the BBC reported user criticism that it amounted to body shaming. TikTok said the filter came from CapCut, a sister company under ByteDance, and that it was blocking videos using it from teen accounts and recommendations.
