OpenAI addresses lawsuit, highlighting challenges of training AI models without copyrighted data.
OpenAI emphasizes the importance of copyright laws in the development of AI technology.
OpenAI commits to operating within legal boundaries and fostering stability within the AI community.
San Francisco, CA – January 10, 2024 – OpenAI, a leading organization in the field of generative AI, has responded to the recent lawsuit filed by The New York Times regarding the use of copyrighted material in training AI models. The company emphasizes the challenges of training AI models without access to copyrighted data and highlights the importance of copyright laws in the development of AI technology.
OpenAI acknowledges the lawsuit and recognizes the significance of creating AI tools like ChatGPT without relying on copyrighted material. However, the OpenAI admits its ‘Impossible’ to train AI without copyrighted data. They explain that copyright law does not forbid the use of copyrighted material for training purposes, as long as it is not used for financial gain.
In their written statement, OpenAI explains, “Because copyright today covers virtually every sort of human expression – including forum posts, government documents, and digital communications – it would be impossible to train today’s leading AI models without using other people’s copyrighted works. However, OpenAI ensures that it would not provide AI systems that infringe upon copyright laws or use copyrighted material without consent.”
The organization further emphasizes that they are committed to meeting the needs of today’s citizens and developing AI systems that comply with copyright regulations. They recognize the value of public domain books and drawings created more than a century ago, which could yield interesting experimental results. OpenAI also highlights the importance of respecting copyright laws and obtaining permission when necessary.
OpenAI’s collaboration with Microsoft is mentioned in relation to their efforts to create AI tools like ChatGPT. They state that without access to copyrighted work, it would be impossible to train large language models effectively. OpenAI cites the House of Lords Communications and Digital Select Committee, which recognizes the challenges posed by copyright infringement when training AI models.
OpenAI concludes their statement by reaffirming their commitment to operating within legal boundaries and fostering stability within the AI community. They maintain that training AI models without copyrighted data remains a complex issue, which necessitates a careful balance between innovation and respecting intellectual property rights.
For media inquiries, please contact: