OpenAI fired back against the New York Times’ bombshell copyright lawsuit over ChatGPT training data. In a new blog post, OpenAI asserts the claims are “without merit” and argues its use of online content is protected “fair use.”
The Times alleges OpenAI and Microsoft stole millions of articles without permission to train their AI models. But OpenAI contends compiling publicly available data to advance technology is legal in the EU, Asia, and beyond.
While OpenAI believes they’re in the right legally, they stress being good citizens matters more. They pioneered easy opt-out systems so publishers can block access, which the Times adopted before suing anyway.
In fact, OpenAI says they were in positive talks with the Times about a ChatGPT collaboration until days before the surprise lawsuit. The proposed partnership would have let Times content appear with attribution in ChatGPT to engage readers.
OpenAI also downplayed the influence of Times articles, claiming their models aren’t substantially shaped by any one source. And the company works to minimize regurgitation, noting the Times wouldn’t provide examples pre-lawsuit.
Responding to OpenAI’s post, the Times’ lawyers reaffirmed the unauthorized use of their journalism to build ChatGPT without payment. In their view, it’s not fair use whatsoever.
Ultimately, OpenAI believes the unexpected lawsuit is disappointing given their efforts to advance AI responsibly and support journalism. They pioneered opt-outs, proposed a Times partnership, and minimized replication risks.
While the Times feels wronged, OpenAI maintains training on available data is lawful and they’ve taken steps to be good citizens. But despite their best intentions, the Times struck back hard.
This battle raises big unsettled questions around AI ethics and copyright law. With two powerhouses locking horns, the gloves are off as OpenAI and the Times wrestle over AI’s massive potential versus authors’ rights.
