OpenAI whistleblower Suchir Balaji found dead; police rule out foul play

The 26-year-old Indian American had accused OpenAI of using copyrighted content to train its AI models. Details

Former artificial intelligence (AI) researcher at OpenAI Suchir Balaji, 26, was found dead in his San Francisco apartment, SiliconValley.com reported on Friday.

 

The Indian-American had worked with OpenAI from November 2020 to August 2024 before becoming a whistleblower, accusing the company of using copyrighted material to train its AI models, including ChatGPT.  

 

Police say death due to suicide

San Francisco police discovered Balaji’s body at around 1:15 pm on November 26 after a wellness check was requested. According to Officer Robert Rueca, a spokesperson for the San Francisco Police Department (SFPD), there was "no evidence of foul play" found during the investigation. The San Francisco Medical Examiner’s Office later confirmed the cause of death as suicide, as reported by The San Francisco Standard.

 

 

Who was Suchir Balaji?

Balaji was a computer scientist who worked at OpenAI for nearly four years.

 

Balaji joined OpenAI in 2020 as an intern and, following his graduation from the University of California, Berkeley, with a degree in computer science, continued at the company until August 2024. He contributed to various projects, including pre-training for GPT-4 and reasoning for o1, as well as post-training for ChatGPT, according to his LinkedIn profile.

 

Accusations against OpenAI

Balaji’s death comes amid ongoing lawsuits against OpenAI and its partner Microsoft. The company faces multiple legal challenges, including allegations of copyright infringement.

 

In an October interview with The New York Times, he voiced concerns about the ethical implications of generative AI, specifically questioning OpenAI’s approach to copyright issues and its use of protected content. Balaji, who had worked at OpenAI for nearly four years, quit the company in August 2024. He told The New York Times that, over time, he realised the technology would likely cause more harm than good to society, particularly due to concerns about OpenAI’s alleged misuse of copyright data.

 

Balaji also expressed his views on copyright issues in a post on X (formerly Twitter) in October, where he revealed how his understanding of copyright evolved. “I initially didn’t know much about copyright, fair use, etc., but became curious after seeing all the lawsuits filed against GenAI companies,” he wrote. “When I tried to understand the issue better, I eventually came to the conclusion that fair use seems like a pretty implausible defence 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.” Balaji also emphasised that this wasn't a critique of ChatGPT or OpenAI specifically, but rather a broader issue related to fair use and generative AI.

 

In the same post, Balaji encouraged machine learning researchers to delve deeper into the subject of copyright, stating that understanding the law and its purpose is crucial, even for non-lawyers. He highlighted that precedents like Google Books were often cited in these discussions, but may not be as supportive as they appear. “Fair use and generative AI is a much broader issue than any one product or company,” Balaji added, inviting others to contact him for further discussion on the intersection of fair use, machine learning, and copyright. 

I recently participated in a NYT story about fair use and generative AI, and why I'm skeptical "fair use" would be a plausible defense for a lot of generative AI products. I also wrote a blog post (https://t.co/xhiVyCk2Vk) about the nitty-gritty details of fair use and why I…

— Suchir Balaji (@suchirbalaji) October 23, 2024