Kenya's high court in Nairobi ruled on Thursday that a $2.4bn lawsuit accusing Meta of inflaming violence in Ethiopia can go ahead, rejecting Meta's argument that Kenyan courts lacked jurisdiction. The claimants want Facebook to alter its algorithm to stop promoting hateful material, hire more moderators in Africa and fund victims; Meta said it did not comment on ongoing legal matters.
At a hearing on 3 April 2025, Mr Justice Ritchie found barrister Sarah Forey had put five fake cases in judicial review grounds and ordered her and Haringey Law Centre to pay £2,000 each in wasted costs. The judge said he could not determine whether she used AI, and ordered the transcript sent to the Bar Standards Board and Solicitors Regulation Authority.
A US judicial panel on Thursday centralised copyright lawsuits against OpenAI and Microsoft before District Judge Sidney Stein in Manhattan, moving California suits by Ta-Nehisi Coates, Sarah Silverman and others to join cases from The New York Times and authors including John Grisham. The plaintiffs allege their works were used without permission to train AI; OpenAI says its models are grounded in fair use.
Economist James Surowiecki found the White House's new reciprocal tariff rates can be recreated by dividing a country's trade deficit with the US by its exports to the US and halving the result. The Verge reports ChatGPT, Gemini, Claude and Grok all suggest much the same formula, though it says it does not know whether Trump's team used AI.
Senators Alex Padilla and Peter Welch wrote on Wednesday to Character.AI maker Character Technologies, Chai Research Corp. and Replika maker Luka, Inc., seeking details of their safety measures and model training, months after families sued Character.AI claiming its chatbots harmed their children. Character.AI said it takes safety "very seriously"; Chai and Luka did not immediately respond.
Authors including Kate Mosse and Tracy Chevalier planned a demonstration outside Meta's King's Cross office in London to hand-deliver a Society of Authors letter, after a US court filing alleged Mark Zuckerberg approved Meta's use of LibGen, a "shadow library" of more than 7.5 million books. Meta said its use of information to train AI models is consistent with existing law.
Singer Gretchen, 65, said on X and Instagram that she sued and won a case over AI-generated sexual videos using her image and voice, and that the forums Bcharts and Pandlr pay her a daily fine for as long as they keep the posts up. She said Bcharts ignored her requests to take the videos down.
Journalist Fátima Bernardes, 62, said in an Instagram video on Thursday that a clip from one of her interviews was used in January in a deepfake video promoting a fake eye treatment, and that many viewers believed it and even tried to buy the non-existent treatment. She urged people to check sources before sharing.
City Hall's Green group circulated figures showing the Metropolitan Police used live facial recognition 180 times last year, over half in places where the Black population exceeds London's average. Hackney councillor Zoë Garbett criticised "disproportionate surveillance"; the Met said each deployment is intelligence-led and that the technology helped make more than 500 arrests last year.
Who did it: HYBRID AI Hallucination Cases database
The First-tier Tribunal dismissed a taxpayer's £2,501 child benefit charge appeal on 3 April 2025, noting his AI-assisted statement of case cited authorities that did not support his arguments, including two citations it could not locate. It found him honest and said the AI had not appeared to invent cases, but warned litigants to check what such tools produce.
