Suchir Balaji (November 21, 1998 – November 26, 2024) was an American artificial intelligence researcher who was found dead one month after accusing OpenAI, his former employer, of violating United States copyright law. His death drew widespread attention because of his purported whistleblower status and claims of foul play made by his parents and others. The San Francisco Police Department investigation, however, found "no evidence of foul play", and the Chief Medical Examiner concluded the death was a suicide.
Early life and education Balaji was born in Florida on November 21, 1998, into an Indian-American family. He grew up in Cupertino, California, where both of his parents worked in the technology sector. He started coding with the children's educational tool Scratch at the age of 11, had built his own computer at the age of 13, and wrote a paper about chip design at the age of 14. Balaji attended Monta Vista High School and was a finalist for the 2015–16 season of the US Computing Olympiad. In 2017, he ranked seventh place in a Kaggle "Passenger Screening Algorithm Challenge" sponsored by the US Transportation Security Administration (TSA), for which he earned $100,000. In 2017, he also won first place in the Pacific Northwest Regional Programming Contest and the Berkeley Programming Contest. In 2018, he placed 31st in the World Finals of the International Collegiate Programming Contest of the Association for Computing Machinery. Balaji took a gap year after high school to work as a software developer at Quora. He then attended the University of California, Berkeley and interned at Scale AI in 2019, and graduated with a BA in computer science. After graduating in 2021, he joined OpenAI full-time.
Career John Schulman, a cofounder of OpenAI, recruited Balaji right out of college. Balaji spent nearly four years working at the company as an artificial intelligence researcher. Among other projects, he was involved in gathering and organizing the internet data used to train GPT-4, a language model used by the company's online chatbot, ChatGPT. He also worked on a precursor model called WebGPT. Writing in an online eulogy, Schulman claimed that "Suchir's contributions to this project were essential, and it wouldn’t have succeeded without him." Balaji left OpenAI in August 2024 after becoming disillusioned with its business practices, saying "If you believe what I believe, you have to just leave the company." After leaving, he said he had been working on "personal projects". According to one interview with his mother along with documents shared by his advisors, he planned to create a nonprofit centered around machine learning and neuroscience that would serve as a direct competitor to OpenAI. While people around him told him to be careful going against his former employer, he remained steadfast in his mission, and worked on getting venture capitalists interested in his project.
New York Times article In an October 23, 2024 New York Times interview, Balaji alleged that products like ChatGPT violate United States copyright law because they are trained on the products of business competitors, and because the chatbots' outputs can then imitate and substitute those products. He said that ChatGPT and similar chatbots are ruining the commercial viability of the individuals and organizations who produced the data that the AI systems are trained on. The New York Times piece contains a summary of Balaji's essay "When does generative AI qualify for fair use?", published on his personal website earlier that week. In the essay, he mathematically analyzed the outputs of chatbots such as ChatGPT, and argued that they fail the four-factor test for determining fair use under U.S. copyright law. He further suggests that the argument could be applied to other generative AI products as well. At the time, OpenAI was being sued for copyright infringement by prominent authors and news publishers, including The New York Times. In a November 18, 2024 court filing, Balaji was identified by the New York Times' attorneys as one of a number of people who might have "relevant documents" in the copyright case against OpenAI. Several of the people named in the news service's court filings as potentially having relevant documents were former or current OpenAI employees. Balaji had said that he would testify against OpenAI. Unlike other OpenAI whistleblowers, Balaji did not reveal any new information about the company. The New York Times article cites Stanford University law professor Mark Lemley, who disagreed that generative AI services violate copyright law, and intellectual property attorney Bradley Hulbert, who said a new law might be necessary to settle the question of legality. Months after Balaji's death, which attracted significant public attention, Hulbert told Fortune magazine that Balaji's essay "[reads like] the argument of a really smart non-lawyer who read up on the subject but does not have a thorough understanding". Another intellectual property attorney, quoted anonymously by Fortune due to concerns about conspiracy theories surrounding Balaji's death, said Balaji's analysis "misunderstands the law in some fundamental ways". OpenAI argued that its software was "grounded in fair use and related international copyright principles that are fair for creators and support innovation", and was "trained on publicly available data".
… excerpt ends here. Continue reading the full article.
