The Author's Guild, a prominent professional organization for writers in the United States, initiated a class-action lawsuit against OpenAI, backed by Microsoft, on September 19. The lawsuit alleges that OpenAI has improperly employed copyrighted material in training its artificial intelligence (AI) models, constituting a violation of the Copyright Act.

According to court documents, The Author's Guild is pursuing legal action to address what it characterizes as "flagrant and harmful infringement" of copyrighted written works of fiction. It claims that OpenAI copied these works in their entirety without obtaining proper permission or offering any form of compensation. The copying was performed by feeding the copyrighted material into large language models (LLMs).

The lawsuit contends that OpenAI's commercial operations heavily rely on these algorithms, which are, in turn, founded on large-scale copyright infringement. The Author's Guild argues that its class of professional fiction writers, whose livelihoods are rooted in their creative literary expression, is put at risk by the unchecked use of LLMs.

The Guild suggests that OpenAI could have chosen to train its AI models using content from the public domain or by paying licensing fees for the copyrighted works. However, it asserts that OpenAI opted to sidestep the Copyright Act entirely, profiting from the unauthorized use of datasets comprised of relatively recent books.

The lawsuit comes on the heels of a previously filed case against Meta and OpenAI, both accused of employing copyrighted material in training their AI models. In this earlier lawsuit, initiated in July, author Sarah Silverman and others sought legal action. However, both companies have subsequently requested the dismissal of the claims.

Additionally, in August, the U.S. Copyright Office initiated a notice of inquiry focused on AI, soliciting public input on topics related to AI content production and its regulation by policymakers. This inquiry delves into how AI-generated content should be handled when it closely resembles human-created content.

Before this inquiry, U.S. District Judge Beryl Howell had ruled that artwork generated solely by AI is ineligible for copyright protection. These legal developments underscore the complex and evolving intersection of AI, copyright, and intellectual property rights, with significant implications for both creators and AI developers.