Encyclopedia Britannica and Merriam-Webster have filed a lawsuit against OpenAI, saying the company used their copyrighted material to train its AI systems without permission.
The complaint, filed on Friday, also claims that OpenAI’s models can produce answers that are very close to Britannica’s original writing.
Britannica says OpenAI copied its content many times and used it in training models like GPT-4. The company argues that GPT-4 appears to have memorized large parts of its articles and can sometimes return passages that are nearly identical to the original text when asked in certain ways.
The lawsuit includes examples comparing responses from OpenAI’s models with Britannica’s published content. According to Britannica, some of these passages match almost word-for-word. The company says this shows its work was copied and then reproduced without authorization.
Britannica also argues that OpenAI is hurting its business by giving users direct answers that replace the need to visit Britannica’s website. It says these AI-generated responses compete with its content instead of sending traffic back to the source, the way search engines usually do.
This case is part of a larger wave of copyright lawsuits brought by publishers and media companies against AI firms. The New York Times has made similar accusations in its ongoing case against OpenAI, claiming that the company copied large amounts of its copyrighted reporting. In another major case, Anthropic settled a class action lawsuit in September over the use of copyrighted books for AI training, leading to a $1.5 billion payout to authors.
If this article helped you, please consider supporting our work. Every small contribution keeps Abijita.com independent and running.
