<\/p>\n
Japan\u2019s largest newspaper is suing Perplexity, accusing the AI search startup of copying more than 119,000 articles to train and power its chatbot. Yomiuri Shimbun Holdings said it filed the claim on August 7 in Tokyo District Court, seeking about \u00a52.17bn ($15m) in damages and an injunction to stop Perplexity from reproducing or distributing its content.<\/p>\n
The suit, brought by three Yomiuri corporate entities in Tokyo, Osaka and Fukuoka, alleges that Perplexity servers accessed the paper\u2019s site between February and June to harvest articles later used in responses to user queries. The company argues this infringes reproduction and public-transmission rights under Japanese copyright law. <\/p>\n
Jiji Press, via Nippon.com, reported the paper is also concerned that \u201czero-click\u201d answers reduce site visits and threaten the business model that supports its journalism.<\/p>\n
Under a 2018 amendment to Japan\u2019s Copyright Act, reproductions for machine learning and data analysis can be made without prior authorisation, but only if they do not \u201cunreasonably prejudice\u201d rights holders. Legal analysts say the Yomiuri case will be an early test of how far that exception extends.<\/p>\n
The action is also tied to web-security claims. Cloudflare has alleged that Perplexity used undeclared crawlers which masked their identity and evaded robots.txt and firewall rules. The company said it removed Perplexity from its verified bot list and updated security rules, with its findings likely to feature in court if Yomiuri pursues that argument.<\/p>\n
Perplexity has grown rapidly in the AI search market, with Bloomberg reporting its valuation reached about $9bn in late 2024. It has millions of users and runs subscription and revenue-sharing programmes with publishers.<\/p>\n
The lawsuit is the first brought by a major Japanese publisher against an AI company. Its outcome could influence how the 2018 copyright exceptions are interpreted, and whether courts demand clearer consent or compensation mechanisms for large-scale crawling that damages news organisations\u2019 commercial interests.<\/p>\n