The New York Times has filed an official complaint against OpenAI and Microsoft, the company behind ChatGPT. The newspaper is considering this step to defend its copyright.

The New York Times had been considering taking legal action against OpenAI for some time. Two sources spoke about this to the independent radio station NRP in August. Today the newspaper also filed an official complaint, writes the Wall Street Journal. Not only against OpenAI, but also against Microsoft.

Conversation ended

Like many online materials, the NYT articles are suitable for training large language models. The newspaper has been in talks with OpenAI for some time with the aim of agreeing a price for the use of its texts. That consultation failed, prompting the Times to take legal action to secure what it believed was fair compensation.

Now the newspaper has adjusted its terms of use. It now states that all of its material is protected by copyright and may not be used to develop software. This also includes training AI.

In a lawsuit in New York

The newspaper expresses major reservations and even sees ChatGPT as a possible competitor in the near future. Especially because the chatbot uses content from NYT articles and interviews in its answers, i.e. things written by people from the newspaper. For example, there are fears that AI-generated answers will perform better than the original in search engines, which are also increasingly driven by AI. Users would then look at the original much less often.

Models like ChatGPT scour the internet as training for whatever information they can find. Such chatbots base their answers and suggestions on this data. If the Times goes through and wins, it would be a major blow to large language models like OpenAI. This would mean nothing less than completely changing their records and only using approved material.

Federal penalties in such cases are high in the United States and can reach up to $150,000 per violation. It is clear that this would spell nothing short of a disaster for AI companies as they are dealing with millions of items being processed. Furthermore, if the NYT perseveres and wins, it could trigger a snowball effect of unprecedented proportions as others undoubtedly follow the paper’s lead.

Not the only ones

There are already organizations and individuals engaged in legal disputes with AI companies. For example, comedian Sarah Silverman joined a class action lawsuit against OpenAI. She claims she never gave the company permission to use her bio, which in fact happened. Fiction and non-fiction authors are no longer willing to simply take their works off the Internet to train models.

It’s not just about LLMs and texts either. Getty Images has now taken Stability AI, the company behind Stable Diffusion, to court. The provider of images never consented to the use of its library for AI training. Getty’s photos can be identified by the semi-transparent frame in his photos when no payment has been made for them, which has already effectively shown up in AI image generators that have trained on it and simply adopted the label.

This article originally appeared on August 18th. The text has been updated with the latest information.