OpenAI, the developer behind the innovative chatbot ChatGPT, has argued that creating advanced AI tools like ChatGPT would be "impossible" without access to copyrighted material. The statement comes in response to increasing scrutiny on AI firms regarding the content used to train their products.
Chatbots and image generators, including ChatGPT and Stable Diffusion, are trained on extensive datasets sourced from the internet, much of which is protected by copyright.
OpenAI's submission to the House of Lords communications and digital select committee highlighted the indispensable role of copyrighted works in training large language models like GPT-4, the underlying technology for ChatGPT. The company emphasized that contemporary AI models rely on a wide range of human expression, from blog posts and photographs to software code and government documents, all covered by copyright.
Addressing concerns raised by legal actions, including a lawsuit by the New York Times against OpenAI and Microsoft, OpenAI stated that limiting training materials to out-of-copyright books and drawings would result in inadequate AI systems. The company argued that such limitations might yield interesting experiments but would not meet the needs of today's citizens.
In response to the NYT lawsuit, OpenAI published a blog post reiterating its support for journalism, collaboration with news organizations, and its belief that the lawsuit is without merit. The company has consistently expressed respect for content creators' rights and emphasized the legal doctrine of "fair use," which allows the use of copyrighted content under certain circumstances without seeking permission.
OpenAI's legal challenges extend beyond the NYT lawsuit, with authors such as John Grisham, Jodi Picoult, and George RR Martin suing the company for alleged "systematic theft on a mass scale." Getty Images is also suing the creator of Stable Diffusion, Stability AI, for alleged copyright breaches, while a group of music publishers, including Universal Music, is suing Anthropic, an Amazon-backed company behind the Claude chatbot.
In its House of Lords submission, OpenAI expressed support for independent analysis of its security measures, endorsing "red-teaming" where third-party researchers emulate the behavior of rogue actors to test AI system safety. OpenAI is also among the companies collaborating with governments on safety testing powerful models, following an agreement reached at a global safety summit in the UK last year.
Comments