Yksi maailman suurimmista kirjojen kustantajista haastoi OpenAI:n oikeuteen tekijänoikeusloukkauksesta
openai
| Source: Mastodon | Original article
Penguin Random House, one of the world’s largest book publishers, has filed a lawsuit against OpenAI, accusing the AI firm of infringing its copyrights by using a German children’s‑book series in the training of ChatGPT and other models without permission. The publisher says the texts were scraped from its catalog and fed into the company’s massive language‑model datasets, enabling the system to reproduce passages and generate derivative content that competes with the original works.
The case spotlights a growing clash between traditional media owners and the rapidly expanding AI industry. As generative models become more capable, they rely on ever‑larger corpora of copyrighted material, often harvested from the public internet. Rights holders argue that such use amounts to wholesale copying that bypasses licensing fees, while AI developers contend that the data is transformed under fair‑use or similar doctrines. Recent rulings in Germany, where the music‑rights collective GEMA successfully sued OpenAI for unlicensed training material, and the pending New York Times suit against the same company, suggest courts are willing to scrutinise the practice.
What follows will likely shape the economics of AI development. If Penguin Random House secures an injunction or damages award, OpenAI may be forced to negotiate blanket licences with publishers, potentially adding significant costs to its pricing model. The outcome could also prompt other content creators—film studios, news outlets, and software firms—to pursue similar actions, accelerating the push for clearer legal frameworks around AI training data. Observers will watch the court’s handling of the German‑book claim, any settlement talks, and whether regulators in the EU or US move to codify data‑use rules before the litigation concludes. The verdict could set a precedent that determines whether generative AI can continue to learn from existing cultural works without explicit permission.
Sources
Back to AIPULSEN