Alec Radford, a former OpenAI researcher instrumental in developing the company’s AI models, has been subpoenaed in a copyright lawsuit filed against OpenAI. According to a court filing, Radford was served the subpoena on February 25 as part of an ongoing case in the U.S. District Court for the Northern District of California.
The lawsuit, brought by authors including Paul Tremblay, Sarah Silverman, and Michael Chabon, claims OpenAI violated copyright laws by using their works to train its AI models without proper attribution. The plaintiffs argue that ChatGPT reproduces portions of their copyrighted material without permission.
Radford, who left OpenAI in late 2023 to focus on independent research, was the lead author of a foundational research paper on generative pre-trained transformers (GPTs), which power ChatGPT and other OpenAI products. During his tenure at OpenAI, he contributed to key AI models, including the GPT series, the Whisper speech recognition system, and the DALL-E image generator.
The lawsuit has faced mixed legal outcomes so far. Last year, the court dismissed two claims against OpenAI but allowed the direct copyright infringement claim to proceed. OpenAI defends its practices by arguing that training AI models on copyrighted content falls under fair use. The plaintiffs’ legal team is also seeking testimony from former OpenAI employees Dario Amodei and Benjamin Mann, who later founded the AI company Anthropic. However, both have resisted the motions, citing excessive legal burdens.
The case highlights growing legal challenges surrounding AI training data and copyright laws, with OpenAI and other tech firms facing increasing scrutiny over how their AI models are trained. The outcome could have significant implications for the future of AI development and copyright protection.
