The New York Times says OpenAI hid evidence in ChatGPT copyright trial

The New York Times and Daily News allege that OpenAI was lying about its ability to search customer chat history data and training datasets for their copyrighted works. It’s the final crescendo for a two-year-old Lawsuit against AI company Allegedly violating copyright law by training its generative AI models on Times content and reproducing that journalism in user output.

Throughout the case, OpenAI argued that it lacked the ability to search its training set. It also argued that searching or producing its massive collection of ChatGPT conversations would be technically burdensome and create a burden Concerns about user privacy Because the records will need to be retrieved, processed and de-identified. The outlets sought this data to determine whether their copyrighted journalism was present in OpenAI’s training dataset, and whether and how often ChatGPT generated responses using or reproducing their content.

In a court-ordered deposition in April, OpenAI’s data privacy engineer Vinny Monaco revealed that OpenAI had already conducted internal searches and evaluations of its training set to look for copyrighted journalistic works.

Monaco’s affidavit also allegedly revealed that, before The New York Times filed its lawsuit, OpenAI had already compiled a database of about 78 million de-identified ChatGPT conversations that it was using internally to determine the extent to which it was infringing on others’ works. On top of this data set, OpenAI also allegedly implemented a so-called “Bloom” filter as part of a set of tools called “Project Giraffe,” which detected and kept a record of regurgitation in the output, shortly after the lawsuit was filed.

These last two revelations are of particular importance. Plaintiffs had originally asked OpenAI to provide a sample of 120 million chat records, but OpenAI negotiated to reduce the sample to just 20 million. OpenAI finally submitted that sample to the courts last December, but it allegedly included too many redactions to make the sample “unusable,” the court said. Plaintiffs also claimed OpenAI Billions deleted Of ChatGPT’s output after they filed a lawsuit in direct violation of the court’s preservation order, the AI ​​giant replaced millions of records in the requested sample.

In other words, they claim that OpenAI made it unnecessarily difficult to obtain information the company had already collected.

“If OpenAI truly believed that copying our clients’ journalism was fair and legal, it would not have hidden the fact of doing so,” Ian B. Crosby, lead counsel for the plaintiffs, said in a statement.

Now, the New York Times and Daily News are asking a judge to discipline OpenAI for allegedly withholding evidence and tampering with the discovery process. They are asking the court to block OpenAI from using the 20 million chat history sample as evidence, claiming it is unreliable; to accept that the ChatGPT logs would have shown significant regurgitation and grounding of the plaintiffs’ content; To prevent OpenAI from saying that the provided chat logs do not show significant regurgitation; And to make OpenAI pay the legal fees for having to chase down this evidence.

In a statement, OpenAI spokesman Drew Pusateri denied the allegations and accused the Times of trying to access user private conversations while its case was weak.

“As the Times’ case weakens and they are forced to drop their claims against us, they are continuing their efforts to violate the privacy of people who have nothing to do with this case, including by making these blatantly false allegations,” Pusateri said. “We will continue to defend our users’ privacy and well-established fair use principles.”