Not even a year has passed since OpenAI’s ChatGPT started spreading like wildfire among Internet users, literally turning not only education upside down, but also many other industries and companies. The language model has been a huge help in many areas of life – but now it looks like it may not stay that way for much longer, as one of the largest media companies in the world is ready to sue OpenAI.
As is Ars Technica Certainly the background to the issue is that the various language models are trained using third party data found on the internet in most cases – OpenAI has also created a separate bot for this purpose. The problem is that in many cases they also used content that was under copyright protection, which of course hasn’t affected companies like OpenAI at all – until now.
me, you, us
Unlike the authors, who seem more concerned with preserving their ability to remove their writing from OpenAI training models, the Times has other concerns about AI tools like ChatGPT. the NPR She stated that a major concern was that ChatGPT could use content from The New York Times to become competitive by creating text based on original reporting and writing by the newspaper’s staff at the request of users.
Among other reasons, the Times has updated its terms of service, which now prohibit its content from being used “for the development of any software, including but not limited to training a machine learning system or artificial intelligence.” The update gives The Times an extra layer of protection, as NPR reports that the outlet appears to be reconsidering its licensing agreement with OpenAI.
This licensing agreement would have ensured that OpenAI would pay for the New York Times content used to train the models, but according to The Information, communication between the two companies quickly became stormy, putting the deal at risk.
They could be in big trouble
If the New York Times team can prove that the company illegally copied and studied their material in order to develop their generative language model, a federal judge could order OpenAI to rebuild its dataset from scratch. In addition, various financial penalties are a huge burden for the AI company, as those violating copyright law are fined up to $150,000 for each material used.
In the case of OpenAI, where there could be tens of thousands of copies, this could easily mean the end.
In this case, we’re already talking about a multibillion-dollar fine, and it’s not certain that OpenAI will be able to fund it — and if The New York Times wins the lawsuit, it becomes almost certain that other media companies will oppose OpenAI, inspired by the case.
Of course, it’s still questionable whether The New York Times will truly abide by its demands and sue OpenAI, but it’s certainly interesting that last month the Associated Press became one of the first news outlets to sign a contract licensing agreement with OpenAI, so it’s certainly Obviously, there are news agencies that want to work with the company. The AP later reported that it had joined other news organizations in developing standards for the use of AI in newsrooms, acknowledging that “many news organizations are concerned about AI companies using their material without permission or payment”.
Follow the index on Facebook too!
