In a landmark legal battle, The New York Times (NYT) has lodged a lawsuit against OpenAI and Microsoft. The charge? Alleged copyright infringement stemming from the unauthorized training of AI models on the newspaper’s content.
- The Contention: The lawsuit, filed in Manhattan’s Federal District Court, alleges that millions of NYT articles were used without permission to train ChatGPT and Copilot, popular AI models owned by OpenAI and Microsoft respectively. The Times’ lawsuit is unique in highlighting potential brand damage through “hallucinations,” or incorrect information generated by AI models.
- The Demand: The newspaper has demanded the obliteration of any models and data containing their copyrighted content, along with billions in statutory and actual damages.
- The Broader Picture: Generative AI models, like the ones in question here, learn from vast amounts of data scraped from the web. This has led to ongoing debates on fair use, copyright infringement, and the rights of copyright holders.
- The Licensing Route: Some news outlets, instead of initiating legal proceedings, have chosen to sign licensing agreements with AI vendors, such as the Associated Press and German publisher Axel Springer.
- Failed Negotiations: Despite attempting to reach a similar licensing agreement, NYT’s negotiations with Microsoft and OpenAI were unsuccessful.
This lawsuit represents a conflict between traditional journalism and technological advancement, carrying significant implications for digital rights in the information age. The outcome could set a crucial precedent for the use of copyrighted material in AI training.