French regulator fines Google €250m for copyright infringement in AI creation

Greg Bulla/Unsplash

France's antitrust regulator has fined Google 250 million euros for violating EU intellectual property rules when working with publishers. The reason was unauthorised scanning of news articles to train artificial intelligence models, including the chatbot Bard (now Gemini).

Here's What We Know

According to the regulator's ruling, Google copied content from news sites to train AI without notifying publishers or authorities. In doing so, publishers were not given the opportunity to opt out of the use of their content, depriving them of their right to fair payment.

The decision is part of a long-running dispute between Google and media companies over the use of online content. In 2019, publications including Agence France-Presse filed complaints against the company for unpaid use of their articles, affecting advertising revenue.

Although Google entered into a news buying agreement with AFP in 2021, the regulator deemed that the company was forcing publishers to go to tricky lengths to make such deals and fined it €500 million in 2022.

The tech giant then promised to make the process of paying publishers more straightforward. However, instead, the company used articles from these media outlets to train Bard/Gemini.

Last year, dozens of publishers accused tech companies of stealing the work of artists and writers to train their artificial intelligence models. In December, the New York Times sued Microsoft and OpenAI in a similar case.

Despite the EU's artificial intelligence law, there are no specific regulations protecting authors from unpaid use of their work to train AI. At the same time, some publishers, such as Axel Springer, have multi-million dollar deals with OpenAI that allow their content to be used in exchange for payment.

Source: The Next Web