AI companies are under a tsunami of lawsuits for the unauthorised use of copyrighted materials to train their AI models. Authors, artists and publishers in the United States of America, France, Australia, Canada, and India have taken legal actions, marking a global reckoning over intellectual property rights in the age of AI. The New York Times has sued Peperlexity, the artificial intelligence start-up that has built a cutting-edge internet search engine, and OpenAI for training their models on NYT content without authorisation.
The largest legal action to date resulted in a $1.5bn settlement, after a coalition of authors sued Anthropic for using nearly 200,000 copyrighted books to train its large language models without consent or compensation. Court filings revealed that a significant portion of the training data used by Anthropic – as well as Meta and Bloomberg – was drawn from pirated book databases such as Books3 and LibGen, exposing just how deeply unauthorised content has been embedded in the development of frontier AI systems.
The authors whose fiction and non-fiction books were used to train these foundational AI models without consent included globally celebrated African writers, such as South African novelist Zakes Mda; Nigerian Nobel laureate Wole Soyinka; the internationally renowned Chinua Achebe; Nigerian author Chimamanda Ngozi Adichie; Moroccan-American writer Laila Lalami; and Ghanaian literary icon Ama Ata Aidoo, among others. Some of these African writers whose publishing rights were managed by houses in the US, UK, and France are in line to be compensated as part of Anthropic’s $1.5bn settlement.
