The New York Times and a coalition of major American newspapers have escalated their copyright battle against OpenAI by petitioning a federal Manhattan court to impose sanctions, alleging that the artificial intelligence firm made false statements under oath regarding its technical capabilities. In a filing submitted on Thursday, the publishers claimed OpenAI deliberately misrepresented whether it could search its large language models to detect evidence that millions of their articles had been improperly used to develop ChatGPT, the company's widely-used generative AI chatbot.
The newspapers contend that OpenAI's representations to the court were fundamentally dishonest. According to their legal filing, OpenAI had told the court that searching its systems for copyrighted material would be technologically infeasible and burdensome, while simultaneously concealing that the company had actually conducted such searches well before the initial lawsuit was filed in 2023. This apparent contradiction forms the crux of the newspapers' request for court-imposed penalties and attorney fees.
In their motion, the publishers also highlighted what they describe as a systematic erasure of evidence. They assert that OpenAI has deleted billions of ChatGPT conversations or rendered them inaccessible, actions the newspapers characterise as an obstruction of their ability to prove the company's misuse of their intellectual property. The filing further requests that the court make a formal finding that OpenAI's chat logs and communications demonstrate the company knowingly exploited copyrighted news content.
The allegations carry significant weight given the testimony of an OpenAI employee who later admitted in court that the company had "performed multiple searches for News Plaintiffs' content," directly contradicting the company's prior assurances to the judiciary. This discrepancy has given the newspapers ammunition to characterise OpenAI's conduct as deliberately deceptive rather than merely technically mistaken.
Ian Crosby, the lead attorney for the New York Times, condemned the company's behaviour in a statement, saying OpenAI had "lied to The Times, The Daily News Plaintiffs, the public, and the court" for more than two years. He detailed how OpenAI had falsely claimed that searching ChatGPT outputs for copies of the newspapers' content would be infeasible, burdensome, and invasive of user privacy, while the company was simultaneously performing exactly such searches.
This dispute represents one front in a broader legal and ethical reckoning over artificial intelligence development practices. Beyond the New York Times and its allied publishers, numerous copyright holders—including authors, visual artists, and music labels—have initiated separate legal actions against technology companies including Anthropic and Meta Platforms, all raising similar allegations that their protected creative works were appropriated without consent to train AI systems.
The case originally commenced when the New York Times filed suit in 2023, accusing both OpenAI and Microsoft, which is OpenAI's largest investor and major financial backer, of systematically utilising millions of the newspaper's articles to train the language model powering ChatGPT. The lawsuit specifically targeted what the Times characterised as unauthorised commercial use of its journalism to create a competing product.
OpenAI has not yet responded publicly to the latest motion seeking sanctions. The company's silence on the allegations of court deception stands in contrast to its general approach of defending its AI development practices as lawful and necessary for technological advancement. The absence of immediate comment may reflect the gravity of the allegations, which go beyond merely disputing whether copyright infringement occurred to questioning the company's truthfulness and transparency with the judicial system.
The outcome of this particular motion on sanctions could significantly influence how the broader case proceeds. If the court agrees with the newspapers' characterisation of OpenAI's conduct as deliberately misleading, it may be inclined to impose penalties that extend beyond the typical remedies in copyright disputes. Such a determination could also set a precedent for how courts treat allegations of deception by AI companies in intellectual property litigation.
For Southeast Asian observers and policymakers, this case illuminates the tensions between AI innovation and intellectual property protection that are increasingly relevant in the region. As companies across Asia develop and deploy AI systems, the fundamental question of how generative AI should be trained—and what uses of existing creative works constitute fair use versus infringement—remains contentious and unresolved across most jurisdictions.
The case also raises important questions about corporate accountability and judicial transparency in the age of AI. The alleged deletion of billions of conversations suggests a potential problem of evidence management and data governance that extends beyond this particular dispute, touching on broader concerns about whether AI companies are adequately documenting and preserving their training processes and operational decisions.
