There have been interesting developments in the court case between The New York Times, Open AI, and Microsoft. In December 2023, the news company sued the tech giants for infringing copyrighted works. New details have surfaced where OpenAI accuses The Times ohacking into the company products.
The lawsuit by the plaintiffs, in this case, The New York Times, argued that these ChatGPT-4 models were using Times’ articles to train their chatbots. In the court filing, Times revealed 100 examples to defend its case.
Open AI argued that the newspaper subscription paid someone to hack into OpenAI products, exploit a bug, and then generate desired responses. The filing added:
“It took them tens of thousands of attempts to generate the highly anomalous results that make up Exhibit J to the Complaint. They were able to do so only by targeting and exploiting a bug (which OpenAI has committed to addressing) by using deceptive prompts that blatantly violate OpenAI’s terms of use. And even then, they had to feed the tool portions of the very articles they sought to elicit verbatim passages of, virtually all of which already appear on multiple public websites.”
Here, Exhibit J refers to the attachments provided by Times to argue against OpenAI. OpenAI also argued that the company is making efforts to assist journalism and that no party has the right to “monopolize facts or the rules of language.”
Interestingly, the filing also revealed that in 2020, when GPT-3 was introduced, The Times enthusiastically reported that the technology would be useful and did not accuse Open AI of any copyright infringements.
Microsoft is also a part of the lawsuit. The December filing by The Times talked about the “Browse by Bing” service by Microsoft and powered by ChatGPT. It suggested that the search engine produced articles identical to those of Wirecutter, a site where The Times posts product reviews.
This led to a fall in revenues for the news reporter since Wirecutter earns commissions once it recommends particular products. However, yesterday’s filing suggested otherwise. OpenAI argued that the articles in the complaint were generated by Times asking ChatGPT to provide specific recommendations from the website and that the AI responded by directing the user to the website and providing a “non-verbatim” summary of the review.
For now, we only have information about what was shared in the court. The involved parties have not shared comments about the filing otherwise.
Via: TorrentFreak