Suchir Balaji, a 26-year-old artificial intelligence researcher and former OpenAI employee, was found dead in his San Francisco apartment, according to the city’s Office of the Chief Medical Examiner. In October, Balaji publicly raised concerns about OpenAI’s copyright practices during an interview with The New York Times.
“The cause of death has been confirmed as suicide,” a spokesperson for the OCME stated in a comment to TechCrunch. “The next of kin have been notified. No further comments or reports will be released at this time.”
Balaji, who worked at OpenAI for nearly four years, left the company after concluding that its technology posed greater risks than benefits to society. His main concern centered around OpenAI’s alleged use of copyrighted data, which he believed harmed the integrity of the internet.
A spokesperson for OpenAI expressed grief over the news: “We are deeply saddened by this tragic loss and extend our heartfelt condolences to Suchir’s loved ones during this difficult time.”
On November 26, San Francisco Police Department officers discovered Balaji’s body at his Buchanan Street apartment in the Lower Haight neighborhood. Officers and medical personnel had responded to a wellness check request. Initial investigations revealed no evidence of foul play, police confirmed.
In an October tweet, Balaji shared his thoughts on the legal challenges faced by generative AI companies, including OpenAI. “I worked at OpenAI for almost four years, including 1.5 years on ChatGPT. My curiosity about copyright laws arose after observing the many lawsuits against GenAI companies. Over time, I concluded that many generative AI applications would struggle to justify fair use, as they directly compete with the data used to train them.”
The San Jose Mercury News was the first outlet to report on his death.
A day before Balaji’s body was discovered, on November 25, his name appeared in a legal filing related to a copyright lawsuit against OpenAI. As part of the proceedings, the company had agreed to review internal records related to the copyright concerns Balaji had recently raised.
While several former OpenAI employees have voiced concerns about the company’s safety culture, Balaji was one of the few to openly criticize how the company used data to train its models. In an October blog post, he argued that ChatGPT’s training data did not fall under fair use. He also acknowledged that similar concerns could apply to other generative AI products.
Before joining OpenAI, Balaji studied computer science at the University of California, Berkeley, where he interned at both OpenAI and Scale AI. He later joined OpenAI full-time, contributing to projects like WebGPT, an optimized version of GPT-3 with web-search capabilities. He also worked on GPT-4’s pre-training team, the reasoning team, and ChatGPT’s post-training team, according to his LinkedIn profile.
Following his passing, former colleagues and figures in the AI community took to social media to express their condolences for his loss.
By: Nestor Castillo, ForAllTechNews Director