AI Models: New York Times Sues Perplexity, Meta Closes License Agreements

The dispute over the use of digital content from media houses in AI models is reaching a new level of escalation. In the US, OpenAI has now been ordered by a court to release 20 million anonymized chat logs. These are to be used to verify whether ChatGPT has reproduced copyrighted content from the New York Times verbatim. The newspaper publisher announced meanwhile that it has now also filed a lawsuit against the AI search engine Perplexity.

The more detailed judicial examination of a chatbot's answers shows parallels to the dispute between the German music rights society GEMA and OpenAI. In the German proceedings, it was recently decided that ChatGPT had violated the copyright of several musicians because it could not be a coincidence that the AI reproduces song lyrics so precisely based on probabilities alone.

Fair use or copyright infringement?

In the case in the US, however, the question is not whether ChatGPT has outputted New York Times news content, but how. Within a certain framework, this could pass as "fair use." This is what AI operator OpenAI is hoping for. If the court, on the other hand, finds that an inappropriate amount has been reproduced, it could become expensive. OpenAI had resisted the release of the chat logs for data protection reasons. Furthermore, the majority is irrelevant. However, the judge in the proceedings found that anonymization sufficiently mitigates the logs. The lawsuit against OpenAI and Microsoft has been ongoing since December 2023.

Perplexity also concerns the use of news articles. The AI search engine uses complete reports to answer user queries. From the plaintiffs' perspective, it is further aggravating that the AI sometimes hallucinates. The incorrect result is then often wrongly attributed to the medium. According to its own statements, the New York Times has tried for 18 months to reach an agreement with Perplexity – but unsuccessfully. A lawsuit by Dow Jones is already pending against Perplexity.

Meta closes license agreements

In total, there are said to be 40 similar lawsuits against AI companies in the US. The media houses are not necessarily concerned with completely preventing AI models from using their content. In many cases, they simply expect AI providers to enter into license agreements. It has now become known that Meta has concluded several new agreements. Media houses including USA Today, CNN, Fox News, and Le Monde have signed, Axios reported.

(mki)