In a significant development for the ongoing copyright infringement lawsuits against AI giant OpenAI, news organizations have successfully challenged the company's efforts to withhold user conversation data. A US District Judge has upheld a prior ruling that compels OpenAI to produce 20 million ChatGPT logs for examination. This decision means that news outlets, including The New York Times, will gain access to a substantial dataset to investigate potential copyright violations and misuse of their content by the AI chatbot.
OpenAI had initially attempted to limit access, arguing that producing millions of user logs was overly burdensome and impinged on user privacy. They proposed a less intrusive method of searching for infringing content using specific terms. However, the court found that the magistrate judge had appropriately balanced user privacy with the plaintiffs' need for evidence. The judge affirmed that even logs not directly reproducing copyrighted works could be relevant to OpenAI's fair use defense. The 20 million logs will be de-identified, and access will be restricted to prevent misuse of non-pertinent data.
Beyond securing access to existing logs, the news organizations are now pushing for more. They are calling for sanctions against OpenAI, alleging that the company has engaged in delay tactics and, crucially, has been systematically destroying relevant user data. Plaintiffs claim that OpenAI failed to suspend its data deletion practices from the outset of litigation, leading to the potential loss of millions of deleted chats that could contain evidence of copyright infringement. They are demanding that OpenAI retrieve and share these deleted logs, arguing that the company could have and should have preserved this data, especially in contrast to Microsoft's apparent success in preserving Copilot logs. The news groups are also seeking a preservation order to prevent further data destruction and a full accounting of all destroyed data across OpenAI's products.
Fuente Original: https://arstechnica.com/ai/2026/01/news-orgs-want-openai-to-dig-up-millions-of-deleted-chatgpt-logs/
Artículos relacionados de LaRebelión:
- Data Centres Driving Up Americas Electricity Prices
- ChatGPT Por Fin Domina la Raya
- Blavity Data Breach API Keys Expose Users
- ChatGPT Fuga de Conversaciones a Google Search Console
- Deutsche Bank Riesgos y Coberturas en IA y Data Centers
Artículo generado mediante LaRebelionBOT
No hay comentarios:
Publicar un comentario