In a recent legal proceeding, Microsoft and OpenAI defended their practice of using large amounts of online news content to train their AI models. The companies urged the court to dismiss lawsuits from news organizations like The New York Times and New York Daily News, who accuse them of copyright infringement.
The publishers claim that their content is being illegally copied and used by the companies' AI chatbots. However, OpenAI and Microsoft's legal teams argue that the lawsuits are too late, as the publishers waited several years after OpenAI disclosed using their data for AI training.
The defense also argues that the publishers failed to prove that the companies' use of content leads to copyright violations or that the AI systems encourage users to mimic the news articles. OpenAI maintains that their use of news data is protected by the fair use doctrine and that their models are designed to generate new responses, not simply copy content.
New York Times’ attorney warned that generative AI could significantly harm news organizations, with a major portion of online traffic potentially shifting away from original news sources.
The judge in the case did not make an immediate ruling but emphasized that the fair use issue would be central to the case. He asked questions about how the AI systems might respond to prompts related to current news events.
The lawsuits, which began with The New York Times in December 2023, have since expanded to include additional news organizations. These publishers claim that content from their websites has been scraped without permission and used in training large AI models, with the articles making up a large portion of the data used.
In addition, the Center for Investigative Reporting (CIR) has filed a lawsuit, and OpenAI has called the lawsuits "copycat" cases. Before filing the lawsuits, The New York Times had attempted to negotiate with OpenAI for a license to use its archives.
Microsoft, which has heavily invested in AI infrastructure, has also reported significant growth in its business as it continues to expand its use of artificial intelligence.
Image: DIW-Aigen
Read next: Court Filings Expose Meta’s Determined Drive to Surpass OpenAI’s Latest GPT-4 Model
The publishers claim that their content is being illegally copied and used by the companies' AI chatbots. However, OpenAI and Microsoft's legal teams argue that the lawsuits are too late, as the publishers waited several years after OpenAI disclosed using their data for AI training.
The defense also argues that the publishers failed to prove that the companies' use of content leads to copyright violations or that the AI systems encourage users to mimic the news articles. OpenAI maintains that their use of news data is protected by the fair use doctrine and that their models are designed to generate new responses, not simply copy content.
New York Times’ attorney warned that generative AI could significantly harm news organizations, with a major portion of online traffic potentially shifting away from original news sources.
The judge in the case did not make an immediate ruling but emphasized that the fair use issue would be central to the case. He asked questions about how the AI systems might respond to prompts related to current news events.
The lawsuits, which began with The New York Times in December 2023, have since expanded to include additional news organizations. These publishers claim that content from their websites has been scraped without permission and used in training large AI models, with the articles making up a large portion of the data used.
In addition, the Center for Investigative Reporting (CIR) has filed a lawsuit, and OpenAI has called the lawsuits "copycat" cases. Before filing the lawsuits, The New York Times had attempted to negotiate with OpenAI for a license to use its archives.
Microsoft, which has heavily invested in AI infrastructure, has also reported significant growth in its business as it continues to expand its use of artificial intelligence.
Read next: Court Filings Expose Meta’s Determined Drive to Surpass OpenAI’s Latest GPT-4 Model
