OpenAI has issued a response to the recent legal action initiated by The New York Times, stating that the media outlet is not presenting the complete story. The lawsuit, filed in December 2023, alleges that AI chatbots like ChatGPT, developed by OpenAI, utilize millions of articles from The New York Times as training data, thus posing competition to the newspaper as a source of reliable information.
In its defense, OpenAI explained that its AI models are trained using publicly available internet materials, which it considers fair use under existing legal precedents. The company underlined the significance of this principle for both creators and innovators in promoting US competitiveness. OpenAI further emphasized that it prioritizes its responsibility as a good citizen over legal rights, offering an opt-out process for publishers to prevent its tools from accessing their sites.
Regarding concerns of AI models memorizing and regurgitating copyrighted materials, OpenAI assured that it has implemented measures to limit inadvertent memorization and prevent such issues in the model’s outputs. The company also expects users to act responsibly and refrain from manipulating the models to regurgitate content, which contradicts the terms of use.
Highlighting the vastness of the training data, OpenAI emphasized that any individual sector, including news, constitutes only a small fraction of the overall data. Therefore, a single data source like The New York Times does not have a significant impact on the model’s intended learning.
OpenAI expressed surprise and disappointment at The New York Times’ lawsuit, as the two parties had been engaging in discussions about collaboration for several months. OpenAI claimed that despite the media outlet mentioning instances of regurgitation, it failed to provide any specific examples, hindering OpenAI’s ability to investigate and address the matter adequately.
While OpenAI considers the lawsuit to be baseless, it remains hopeful for a constructive partnership with The New York Times, acknowledging the newspaper’s longstanding history. This legal action by The New York Times sheds light on broader copyright concerns, as various language models reportedly employ training datasets containing copyrighted content from news and digital media organizations, raising concerns among authors and content creators.
In conclusion, OpenAI’s response clarifies its position on fair use of data and addresses the allegations made by The New York Times. The lawsuit and ensuing discussions also highlight the ongoing challenges surrounding copyright and AI technologies.
The source of the article is from the blog regiozottegem.be