Writing the Payout

Tech progress should not come at the cost of creators’ rights

A big legal battle in the United States has made history. A judge approved a $1.5 billion settlement in a case against AI company Anthropic. Authors and publishers said the company used copyrighted books to train its AI without permission. It is the first major settlement of its kind, and it could affect other AI companies such as OpenAI, Microsoft and Meta.

For many years, tech companies collected huge amounts of writing, art and research from the internet. Most creators were never asked for permission or paid. When AI tools became popular, people realised just how much information had been taken. Last week, famous publishers like Hachette and Elsevier also sued Google, saying books from Google Books were secretly used to train its Gemini AI.

The Anthropic case is an important step because it reminds companies that creators have rights. The court said that training AI on creative work can sometimes be legal, but it also ruled that Anthropic broke the law by storing millions of pirated books from illegal online libraries. About 500,000 books are covered by the settlement, and authors and publishers will receive about $3,000 for each book, shared between them. Compared with the billions earned by AI companies, that is not a huge amount.

Many AI companies now sign agreements to pay publishers and news organisations for using their work in the future. But many questions are still unanswered. What counts as “fair use”? When does AI training break copyright laws? Does using pirated books encourage piracy? Courts have not fully settled these issues yet.

The Anthropic settlement is not the end of the debate. But it is an important first step towards making sure that new technology respects the people whose ideas and creativity help make it possible.

END OF ARTICLE