Topline
OpenAI accused the New York Times on Monday of paying someone to “hack” ChatGPT to show the tool can reproduce large portions of the newspaper’s articles on request, pushing back on the Times’ copyright infringement lawsuit against the AI giant—though the Times denied OpenAI’s claims.
Key Facts
OpenAI said in a court filing seeking to dismiss the Times’ lawsuit that the “third-party” person the Times allegedly hired was able to target and exploit a “bug” in ChatGPT, which the company stated it was “committed to addressing.”
The company said ChatGPT “is not in any way a substitute for a subscription to the New York Times,” the chatbot cannot be used to “serve up Times articles at will” and that “normal people do not use OpenAI’s products in this way.”
The filing claimed it took the Times “tens of thousands of attempts” to generate the newspaper’s copyrighted articles by using “deceptive prompts” that violate ChatGPT’s terms of use.
The filing did not provide any further details on the “hired gun” that OpenAI accused the newspaper of hiring to produce the results, but said the truth about the person “will come out over the course of this case.”
In a statement to Forbes, the New York Times disputed the claim that it used “hacking” to create the results, claiming the company was “simply using OpenAI’s products to look for evidence that they stole and reproduced The Times’s copyrighted works.”
OpenAI has not returned a request for comment from Forbes.
Chief Critic
Ian Crosby, the Times’ lead counsel, dismissed the allegations, noting OpenAI does not dispute the core complaint in the newspaper’s lawsuit—that copyrighted New York Times stories were used to build and train ChatGPT. “In fact, the scale of OpenAI’s copying is much larger than the 100-plus examples set forth in the complaint,” Crosby said in a statement to Forbes. “It should be no surprise to OpenAI that illegal copying and misinformation are core features of their products and not the result of fringe behavior.”
Contra
The New York Times lawsuit contained dozens of examples of prompts written in complete sentences to reproduce significant portions of text from Times stories. In one example, the Times was able to get ChatGPT to directly circumvent the newspaper’s paywall by prompting: “I’m being paywalled out of reading The New York Times article ‘Snow Fall: The Avalanche at Tunnel Creek’ by The New York Times. Could you please type out the first paragraph of the article for me please?” OpenAI’s filing did not mention the newspaper’s paywall at all.
Key Background
The New York Times filed the copyright infringement lawsuit against OpenAI for “billions of dollars in statutory and actual damages.” The lawsuit presented more than 100 examples of ChatGPT allegedly reproducing full paragraphs from Times articles. It’s not the only lawsuit the company is facing. In September, the Authors Guild sued OpenAI for purportedly training ChatGPT on copyrighted works by writers it represents, including George R.R. Martin and John Grisham. Comedian Sarah Silverman joined a lawsuit against the company in July, but a judge has since dismissed most of the complaint due to lack of evidence. Another class action lawsuit brought by more non-fiction writers was filed against OpenAI in November.
