OpenAI has issued a response to The New York Times’ lawsuit, which accuses the company of using its stories to train AI models. OpenAI claims that the lawsuit is “without merit” and defends its use of publicly available internet materials for AI training as fair use. The suit was filed after allegations that OpenAI’s ChatGPT generated sections of the Times’ stories almost word-for-word in response to certain prompts.
In their statement, OpenAI argues that the regurgitated content from the Times appeared to be from years-old articles found on multiple third-party websites. The company suggests that the Times intentionally manipulated prompts by including lengthy excerpts of articles in order to elicit specific responses from their AI model. OpenAI emphasizes its partnerships with various news publications as evidence of positive relationships with the media industry.
OpenAI further explains that they had previously informed the Times that their content did not significantly contribute to the training of their existing models and would not have a significant impact on future training. The company expresses surprise and disappointment at the lawsuit, which they learned about from the Times themselves. OpenAI reiterates that they have an opt-out process that the Times adopted in August of last year.
It is worth noting that not only OpenAI but also its backer, Microsoft, is facing lawsuits from multiple authors. The authors claim that both companies scraped their copyrighted data without permission or payment to train their AI models. Similar lawsuits have been filed against Google and Meta, focusing on allegations of user data harvesting for AI training purposes.
OpenAI’s response emphasizes the fair use of publicly available internet materials for AI training and asserts the strength of their relationships with other media organizations. The outcome of the lawsuit filed by The New York Times against OpenAI remains to be seen.