Google fined $270m. Crime? Using data to train Gemini without permission from French new outlet

Google was fined €250 million (US$271 million) by France’s competition watchdog, Autorité de la Concurrence, for failing to reach agreements with news outlets regarding the publication of links to their content. The authority alleged that Google used press articles to train its artificial intelligence technology without notifying publishers or the regulator.

This is not the first time something like this has happened with Google; in 2020, Google received a €500 million fine from the same authority for similar concerns.

The Autorité de la Concurrence claims that Google did not negotiate in good faith with news publishers concerning compensation for the use of their content through linked articles. The regulator also alleges that Google utilized press content to train Gemini.

In a blog post responding to the fine, Google stated their belief that the penalty is disproportionate and does not adequately reflect their efforts to address the raised concerns.

This is not the first time there has been a feud between a news outlet and LLM creators. OpenAI has faced lawsuits and allegations from The New York Times and TIME Magazine. Moreover, since we’re in France, Microsoft’s partnership with Mistral raised eyebrows among some EU lawmakers, who question whether France lobbied for looser AI regulations to benefit Mistral and Microsoft. The French government had denied any prior knowledge of the partnership.

The French action follows a previous investigation in Germany concerning Google’s News Showcase service, which was ultimately resolved through a settlement. In 2014, Google withdrew its News service from Spain in response to a law granting Spanish publications the right to charge for snippets of their work. The service was eventually reinstated in Spain.

More here.