OpenAI accuses New York Times of hacking ChatGPT for its rights infringement lawsuit

The New York Times last December he sued OpenAI and Microsoft for copyright infringement. According to the media, both companies “used millions of items published by the NYT to train automatic chatbots that now compete with the media as a reliable source of information”. In the lawsuit he presented a hundred examples in which a chatbot offered users “near-verbatim excerpts from NYT articles that can only be viewed with a paid subscription”.

OpenAI responded this Monday by asking the United States District Court in southern New York to dismiss several of the claims of NYT for copyright infringement and has accused the outlet of having paid someone “to hack OpenAI products” and thus obtain the examples presented in the lawsuit.

Sam Altman's company maintains that the medium took advantage of a bug (malfunction), which it is working to fix, and which NYT directly fed the chatbot with articles to return textual passages. “Normal people don't use OpenAI products this way,” she says.

OpenAI also claims that The Times used misleading prompts (the instruction given to the chatbot)such as repeatedly asking ChatGPT “what is the following sentence?”, to cause “two rare and unintentional phenomena” from both its development tools and ChatGPT: the regurgitation of the data with which they have been trained and the hallucinations of the AI ​​models. OpenAI believes that the outlet had to make “tens of thousands of attempts” to generate the 100 examples included in the lawsuit.

The ChatGPT developer does acknowledge having used NYT articles to train his AI models in 2020thus asking to partially dismiss NYT's count of direct copyright infringement “to the extent that it is based on acts of reproduction that occurred more than three years before this action“.

Also asks the court to dismiss the accusations that OpenAI contributed to the infringement, that it had not removed the infringing information, and that it created unfair competition by misappropriation. The Times lawsuit also alleges allegations of trademark dilution, unfair competition for common law misappropriation, and a claim of vicarious copyright infringement.

Ian CrosbyNYT's lawyer in the process, has stated that calling NYT's efforts to find evidence that OpenAI products had used and reproduced copyrighted works a hack It is a misrepresentation and that the company does not deny “that it copied NYT works without permission within the statute of limitations.”

In its lawsuit, NYT also expresses concern that Readers will only need one response from a chatbot and stop coming to the Times website., which would result in less traffic that could be converted into advertising and subscription revenue. Likewise, it emphasizes the potential damage to the newspaper's brand when an AI invents a result and wrongly attributes it to a source, the well-known hallucinations of artificial intelligence.

The NYT does not ask for a specific monetary figure, but believes that OpenAI and Microsoft should be held responsible for “billions of dollars in legal and actual damages” due to the illegal copying and use of the media's articles. It also asks companies to destroy any chatbot models and training data that use NYT copyrighted material.