Journalists sue OpenAI for using their articles to train ChatGPT

Since November 2022, the month in which ChatGPT was launched, the system has been based on information published on the web to generate its responses to users. When it comes to questions about current events, the system is usually based on news fragments.

However, the CIR points out that “when they filled their training courses with journalistic works, the defendants had a choice: to respect these works or not, and they chose the latter.

Monika Bauerlein, CEO and spokesperson for the organization in this case, noted that OpenAI and Microsoft have absorbed their stories to “make their product more powerful, but they never asked permission or offered compensation, unlike other organizations license our material.”

He also noted that the chatbot is programmed not to recognize or respect copyright and that statutory damages are at least $750 per infringed work and $2,500 for violation of the Copyright Act.

The company responded to this case by saying that it continues to work collaboratively with the news industry, through partnerships with media outlets to display their content in their products, including summaries, quotes and attributions to direct users to articles. originals.

“One component of the partnerships is the ability to leverage publisher content through various machine learning and training techniques to help us optimize the display of that content and make it more useful to users,” the company said.

Proof of this is the recent alliance that the company announced with Time magazine, through which it will be able to access current and archival articles from more than 100 years of the medium’s history for “several years.”

So far, OpenAI has signed different contracts with media companies, such as the New York Post, The Wall Street Journal, Vox Media, the Financial Times or the Associated Press, as well as discussion platforms, such as Reddit, to train their AI models and make them both more precise in the information they give and more natural in the language they use.