OpenAI Faces New Allegations in Copyright Infringement Lawsuit With Major Media Outlets

The New York Times and The Daily News are intensifying their legal battle against OpenAI, alleging misconduct in handling copyright concerns. The lawsuit, ongoing for two years, accuses OpenAI of illegally training its models on copyrighted journalism. Key points include:

• Allegations that OpenAI misled about its ability to search its training corpus and user data.
• Vinnie Monaco, an OpenAI data privacy engineer, disclosed that internal searches were conducted to assess the presence of copyrighted works.
• Discovery of a 78-million-record ChatGPT conversation database used internally by OpenAI to gauge infringement levels.
• Implementation of a “Bloom” filter by OpenAI to track regurgitated content following the lawsuit initiation.
• Plaintiffs’ demand for a larger data sample, initially set at 120 million logs but reduced to 20 million through negotiation, with claims of excessive redactions rendering the sample nearly unfit for analysis.
• Accusations that OpenAI deleted crucial ChatGPT outputs, violating court orders.

The plaintiffs are urging the court to penalize OpenAI for allegedly concealing evidence and to impose sanctions. They seek to prevent OpenAI from using the supposedly unreliable chat logs in court and want legal costs reimbursed.

OpenAI spokesperson, Drew Pusateri, refutes the allegations, accusing the media outlets of infringing on user privacy as their case loses strength. The spokesperson asserts the company’s commitment to protecting user privacy and adhering to fair use principles.