On 6 June, OpenAI issued a statement announcing that it would appeal the New York Times and its co-plaintiffs ‘ motions. The action required OpenAI to retain indefinitely consumer ChatGPT and API client data, including deleted chat records and sensitive log data.
In a statement, Brad Lightcap, Chief Operator of OpenAI, stated: “We firmly believe that this is an ultra vires act of The New York Times. We are appealing against this order so that the trust and privacy of the user can continue to be paramount.”
In December 2024, the New York Times joined other media to sue OpenAI and Microsoft for illegal use of the New York Times content to train large language models such as ChatGPT and Bing Chat. The New York Times claims that this violates its copyright and threatens the business model of the original innovation. The newspaper indicated last month that as users clear chat records, potential evidence of copyright infringement may be removed.
In May 2025, the Federal Court granted the request, ordering OpenAI not to delete relevant data and to store all the output logs in isolation. According to OpenAI, this requirement not only increases the operational burden, but also conflicts with the company ‘ s commitment to the privacy of the user, which may undermine the trust of the user.
The central dispute in this case is whether the use of copyright-protected materials to train in the production of AI models constitutes “reasonable use”. The New York Times alleges that OpenAI ‘ s tools sometimes generate output that is close to a verbatim copy of its articles and that it can circumvent its payment wall through the extracts generated by AI.
Both sides maintain that they stand at a high moral level. According to The New York Times, it is protecting the press and its ability to work and receive remuneration. Chief Executive Officer Sam Altman of OpenAI accused the media of being “on the wrong side of history”, while OpenAI stated that the data used by The New York Times in the proceedings were “selective”.
With the expansion of the generating AI industry, the courts are becoming a key battleground for data, privacy and intellectual property. This action was one of several high-profile copyright claims filed against OpenAI and other AI companies. In April this year, Ziff Davis, a company with media such as PCMag and Mashable, sued Openi for using its content without consent.
This week, Reddit sued another AI company, Anthropic, for unauthorized access to Reddit data. Anthropic also faces lawsuits from music publishers and authors.
OpenAI ‘ s appeal will be at the hearing stage by 2025, and the court will assess the balance between the need for data retention and privacy protection. If appeals fail, OpenAI may need to permanently adjust its data storage policy to increase compliance costs. If successful, a precedent for data privacy might be set for the AI industry. Whatever the outcome, the case will have a far-reaching impact on the legitimacy of AI training data and the protection of user privacy.
