OpenAI has released a strong statement in response to a lawsuit filed by the New York Times, asserting that the newspaper is not presenting the complete story. The lawsuit alleges that OpenAI, in collaboration with Microsoft, used copyrighted articles from the Times to train its popular ChatGPT chatbot and other AI features. The Times pointed out instances where the chatbot reproduced substantial portions of text directly from their stories.
In its blog post, OpenAI acknowledged the occurrence of such “regurgitation” but emphasized that it is a rare bug that the company is actively working to eliminate. OpenAI also suggested that the Times may have intentionally manipulated prompts and cherry-picked examples to support their claim.
Responding to OpenAI’s statement, Ian Crosby, a lawyer for the New York Times, highlighted that OpenAI openly admitted to using the Times’ work, among other sources, to develop ChatGPT. Crosby argued that OpenAI’s use of the Times’ journalistic content without permission or payment cannot be considered fair use.
Generative AI technology, like OpenAI’s chatbot, relies on large language models that analyze vast amounts of digital text to generate new content. While companies and researchers have long used online data for such purposes, concerns regarding compensation for content creators have led to an increase in lawsuits.
OpenAI acknowledged that there can be instances where the systems memorize text, referring to it as a “rare failure” in the learning process that they are actively working to address.
As debates surrounding AI and intellectual property continue, it becomes essential to strike a balance between innovation and fair compensation for creators. Both OpenAI and the New York Times are likely to navigate the legal proceedings as the industry grapples with the implications of AI technology on copyright infringement.
The source of the article is from the blog rugbynews.at