French watchdog imposes a €250 million fine on Google for AI training practices

Google has been fined €250 million for violating EU intellectual property regulations by utilizing content to train its AI platform without proper notification to publishers and news agencies. The Autorité de la concurrence, previously known as Bard and now rebranded as Gemini, trained its AI model on content without consent, leading to a copyright dispute triggered by complaints from local news organizations.

Google, opting not to contest the facts as part of settlement proceedings, dropped its appeal against an initial €500 million fine issued in 2022. However, the watchdog reopened the case due to concerns over the AI program.

In its statement on Wednesday, the regulator revealed that Google breached four out of seven commitments agreed upon in the settlement, including negotiating with publishers in good faith and transparently sourcing content. The AI program, Bard, launched in 2023, was trained on data from unspecified sources without consent, impeding publishers’ ability to negotiate fair compensation.

The regulator highlighted that Google linked the use of content by its AI service to the display of protected content, further exacerbating the infringement of intellectual property rights.