OpenAI has been battling publishers in courts over whether it broke copyright law in training its AI models, but these investigations have now taken an unexpected turn.
Suchir Balaji, a former OpenAI employee who had accused the company of breaking copyright law, has been found dead in his San Francisco apartment. His body was discovered by authorities in his flat on 26th November after they’d responded to a welfare check at his residence. While the medical examiner has not disclosed the cause of death, police indicated there is “currently no evidence of foul play.” Balaji was 26 years old.
Balaji had graduated from University of Berkley with a BA in Computer Science in 2021. He had worked with Quora for a year, and interned at Scale AI and Helia. He had also interned at OpenAI, and had worked full-time for the company from November 2020 to August 2024 as a Member of Technical Staff.
Three months ago, Suchir Balaji had appeared in a New York Times article, in which he’d accused OpenAI of breaking copyright law. “I was at OpenAI for nearly 4 years and worked on ChatGPT for the last 1.5 of them. I initially didn’t know much about copyright, fair use, etc. but became curious after seeing all the lawsuits filed against GenAI companies. When I tried to understand the issue better, I eventually came to the conclusion that fair use seems like a pretty implausible defense for a lot of generative AI products, for the basic reason that they can create substitutes that compete with the data they’re trained on,” he had posted on X.
Balaji’s picture had appeared as the cover image of the New York Times article titled “Former OpenAI Researcher Says the Company Broke Copyright Law”. He had said that he had proactively reached out to New York Times for the article, which had been widely discussed and shared. In the article, Balaji had contended that OpenAI was negatively impacting businesses and entrepreneurs whose information was utilised to train ChatGPT. “If you believe what I believe, you have to just leave (OpenAI),” he told New York Times, adding that “this is not a sustainable model for the internet ecosystem as a whole.”
Balaji had given several reasons why he felt that models like ChatGPT had broken copyright law in using data from publications such as New York Times to train their AI. He’d said that ChatGPT had caused usage to drop for products such as education-helper site Chegg and programming forum Stack Overflow, which indicated that ChatGPT was impacting the market for products whose data it was likely trained on. He’d also said that while the outputs for ChatGPT seemed novel, they were often interspersed with large amounts of copyrighted work.
Over the last few quarters, OpenAI has been battling lawsuits from publishers including New York Times for allegedly using their copyrighted data to train its models. In India, ANI has similarly sued OpenAI for using its work to train models without permission. OpenAI has consistently denied charges, saying its usage of copyrighted materials amounts to fair use, but with a prominent whistleblower who’d accused OpenAI of breaking copyright law being founded dead in his apartment, there will likely be renewed focus on these lawsuits in the coming weeks.