Recently, Alec Radford, a former OpenAI researcher, was summoned for a copyright lawsuit against the company. Radford received a subpoena on February 25, according to court documents filed Tuesday by the North District District Court of California. The researcher, who had a key role at OpenAI, left the company late last year to focus on independent research.
Radford is one of the leading developers of OpenAI Generative Pretrained Transformer (GPT) technology, a core foundation of OpenAI’s popular AI chatbot platform, ChatGPT. Since joining OpenAI in 2016, he has been deeply involved in the research and development of a variety of GPT series models, and has also made important contributions to the development of the speech recognition model Whisper and the image generation model DALL-E.
The copyright lawsuit, titled "A lawsuit on OpenAI ChatGPT", was filed by several well-known book authors, including Paul Tremblay, Sarah Silverman and Michael Chabon. The plaintiff accused OpenAI of using their work without authorization when training its AI models and claimed that ChatGPT cited their work when generating content, without noting the source.
Although the court dismissed two plaintiffs’ claims against OpenAI last year, the allegations of direct infringement were allowed to continue. OpenAI insists that its use of copyrighted data for model training falls within the scope of reasonable use and does not constitute infringement.
It is worth noting that Radford is not the only former OpenAI employee to be summoned in the case. The plaintiff's lawyers also tried to force the summons of two other former OpenAI employees, Dario Amodei and Benjamin Mann. These two researchers have attracted much attention for founding Anthropic after leaving OpenAI. Amod and Mann objected to the summons, believing the demands were too heavy and unreasonable.
This week, a U.S. District Judge ruled that Ammod had to undergo hours of inquiry that would involve his work during OpenAI, covering two copyright cases, including a lawsuit filed by the Authors Association. This ruling further exacerbates the complexity of the case and also triggers widespread discussions in the industry on copyright issues in AI model training.