US comedian Sarah Silverman and a pair of authors sued Facebook-owner Meta and OpenAI for allegedly using their content without consent to train artificial intelligence language models.

While large language models (LLMs) like OpenAI’s ChatGPT have taken the world by storm for their user-friendly conversational platforms that provide rich results from simple prompts, such platforms have also drawn significant criticism for how they are trained. Platforms like ChatGPT routinely scrape the internet to train their models, and these lawsuits allege that this includes copyrighted content.

The class action lawsuits from Primetime Emmy-winning Silverman along with authors Richard Kadrey and Christopher Golden accuses the companies of using their books without authorisation to develop their LLMs.

In particular, the lawsuit against Meta alleges that leaked information about the Facebook company’s AI business shows their work was used without permission. The suit against OpenAI meanwhile alleges that summaries of the plaintiffs’ work generated by ChatGPT indicate that it was trained on their content.

Both suits seek unspecified money damages on behalf of a nationwide class of copyright owners whose works were allegedly infringed.

The lawsuits echo wider anxieties from writers around the unregulated use of AI, with the tech’s deployment forming a key part of the ongoing Hollywood writers’ strike. On the inverse, the lawsuits also highlight the legal risks facing developers of chat bots who use copyrighted material – such as books and scripts for movies and TV shows – to train their models.

The lawsuit alleges: “The summaries get some details wrong” but still show that ChatGPT “retains knowledge of particular works in the training dataset.”

Share Story: