OpenAI can't seem to get out of trouble as it faces one lawsuit to another. Accused of using copyrighted data without permission, several other news companies are going after the AI startup for infringement.
News Organizations Sue OpenAI
On top of being sued by The New York Times, OpenAI is facing legal action from eight newspaper organizations under one parent company, Alden Capital Group. The claims say that OpenAI and Microsoft used the plaintiffs' content without compensation or consent.
The publications mentioned, New York Daily News, Chicago Tribune, Orlando Sentinel, San Jose Mercury News, and four more, already have evidence at the ready in the form of excerpts taken from ChatGPT conversations, as reported by The Verge.
The same goes for Copilot, which is developed by the company that also backs OpenAI, Microsoft. The chatbot was able to generate news articles verbatim just a day after it was published without citing their sources.
It shows that the companies developing AI models are using copyrighted works as training data. As if that wasn't bad enough, the chatbots are said to be hallucinating facts as well, meaning they are made up and false, and are being attributed to the news publications.
"This lawsuit is not a battle between new technology and old technology. It is not a battle between a thriving industry and an industry in transition. It is most surely not a battle to resolve the phalanx of social, political, moral, and economic issues that GenAI raises."
The complaint also said that it was about how the tech giant and the AI startup are not entitled to use copyrighted content from the newspaper organizations, as they use it to "build their new trillion-dollar enterprises without paying for that content."
OpenAI Did Right by The Financial Times
The AI company is not on the bad side of every news publication after all. The Financial Times, for instance, managed to strike a deal with OpenAI, wherein the latter would use articles from the news outlet to train its AI models in exchange for compensation.
Given that The Financial Times (FT) is a large news organization, OpenAI will be able to train its models with current events and other cultural topics. In fact, the news organization is also set to get more than just compensation for the training data.
According to Engadget, FT also intends to add AI features to its services. Of course, it was assured that the news company is still "committed to human journalism, which could mean that it does not intend to create news pieces with the use of generative AI.
As for how much FT will earn in the deal, the amount to be paid for licensing its data was not disclosed. However, the speculated costs will be around $1 million to $5 million a year, which is a steady stream of revenue aside from subscription fees from readers.
This sets a good example of how training data should be handled. Ironically, OpenAI seems to also be the company that needs to learn that lesson if infringement claims are to be believed. Regardless, it's a good start and could reach smaller creators whose works were used as well.
Related: OpenAI Lands a Deal With The Financial Times for Training Data
