The data, which was stored on a virtual machine provided by OpenAI, contained search results of the publishers’ content within the company’s AI training datasets. This deletion has marred proof for the case.

The incident occurred on November 14, when OpenAI engineers accidentally erased the search data, which had been collected over 150 hours of work since November 1. 

Although OpenAI attempted to recover the lost data, it was unable to restore the folder structure and file names, making the recovered data unusable for identifying where the alleged copyrighted content was incorporated into the training models.

As a result, The New York Times and Daily News’ legal teams now face the task of recreating their work from scratch, requiring further hours of labour and additional computer processing time. 

The publishers’ attorneys have stressed that, while they do not believe the deletion was intentional, the event reveals OpenAI’s position as the entity most capable of conducting the necessary searches using its own systems.

The publishers argue that OpenAI’s use of their content without permission to train AI models like GPT-4 constitutes a violation of copyright law. 

OpenAI, however, defends its actions, asserting that training models on publicly available data is covered under fair use, and that it is not required to pay for or license content used in this manner.

Nevertheless, OpenAI has recently entered into licensing agreements with several publishers, including the Associated Press and the Financial Times, though the terms of these deals remain undisclosed. 

Reports disclose that some of these partnerships could be worth millions of dollars annually, with Dotdash Meredith, one of the partners, reportedly receiving at least $16 million a year.

OpenAI has yet to comment on the specific circumstances of the data deletion or confirm whether its AI systems were trained on the publishers’ content without consent.

OpenAI has shed light on its business practices summarising in four key points, each aimed at clarifying OpenAI’s position and responsibility to ethical AI development.

The company claims to have met with dozens of news organisations and industry leaders to explore opportunities, address concerns, and provide solutions. The goal is to facilitate a healthy news ecosystem, be a valuable partner, and create mutually beneficial opportunities. OpenAI cites partnerships with the Associated Press, Axel Springer, American Journalism Project, and NYU as examples of its efforts to support the news industry.

While asserting that training AI models using publicly available internet materials is fair use, OpenAI acknowledged the importance of respecting creators. The company asserts to have led the AI industry in providing a simple opt-out process for publishers, allowing them to prevent OpenAI’s tools from accessing their sites. OpenAI points out that this opt-out feature was adopted by The New York Times in August 2023.

OpenAI also noted the occurrence of “regurgitation” as a rare failure in the learning process. The company affirmed continually making progress to drive this issue to zero. Measures are in place to limit inadvertent memorisation and prevent regurgitation in model outputs. OpenAI expects users to act responsibly, underlining that manipulating models to regurgitate is not an appropriate use of the technology and is against its terms of use.

OpenAI explained that discussions with The New York Times were progressing constructively until the unexpected lawsuit on December 27. The negotiations were centered around a high-value partnership for real-time display with attribution in ChatGPT. 

OpenAI contends that The New York Times’ content didn’t significantly contribute to the training of existing models, and their lawsuit came as a surprise. The company conveyed disappointment in The New York Times’ refusal to share examples of regurgitation, despite commitments to investigate and fix any issues.

OpenAI challenges the merit of The New York Times’ lawsuit, expressing hope for a constructive partnership and reiterating its obligation to collaborate with news organisations. The response points to OpenAI’s dedication to responsible AI development and the continual improvement of systems to resist adversarial attacks.

The lawsuit, filed in the U.S. District Court for the Southern District of New York, comes at a time when the ongoing debate over the unauthorised use of published content to train artificial intelligence (AI) models is on.

The NYT’s lawsuit accused Microsoft and OpenAI of engaging in a business model founded on “mass copyright infringement.” The heart of the matter lies in the alleged unlawful copying and utilisation of The Times’s substantial body of work to train large language models, including ChatGPT. 

The lawsuit seeks “billions of dollars in statutory and actual damages” for the claimed infringement and emphasises the creation of AI systems that exploit and, in many cases, retain significant portions of copyrighted expression from NYT’s articles.

Microsoft, a key player in the lawsuit,  invests in OpenAI and also supplies the company with access to its Azure cloud computing technology. This close relationship adds complexity to the legal dispute, raising questions about shared responsibilities and obligations.

The lawsuit points to a potential turning point in defining the legal boundaries surrounding generative AI technologies. As AI firms, including OpenAI, attract billions in funding, concerns about the uncompensated use of intellectual property have grown within creative industries. This case could set a precedent for how media organisations navigate agreements and protect their content at a time where AI models can replicate and reimagine written works.

OpenAI, currently valued at over $80 billion, has garnered significant attention and financial support, with Microsoft committing $13 billion to the company. The lawsuit implies that OpenAI and Microsoft, by utilisng The Times’s content without proper authorisation, aim to benefit from the newspaper’s substantial journalistic investment without due compensation.

The NYT’s legal action suggests that ChatGPT and other AI systems created by OpenAI and Microsoft could emerge as competitors in the news industry. The concern is that AI-generated responses, utilising copyrighted material from The Times, may divert readers away from the newspaper’s website, impacting advertising and subscription revenue.

The lawsuit follows unsuccessful negotiations between The Times, Microsoft, and OpenAI. The newspaper had reportedly approached the tech companies in April to address concerns about intellectual property use, seeking an amicable resolution involving potential commercial agreements and technological safeguards around AI products. However, these discussions reportedly reached an impasse, leading to the legal action.