Google France

France Fines Google for $272 Million Over Generative AI Training Practices

France’s competition authority has imposed a fine of $272 million on Google for allegedly failing to comply with commitments made last year regarding the use of news content in training its artificial intelligence systems. The Autorité de la concurrence stated that Alphabet, Google’s parent company, breached its commitment to cooperate with a monitoring trustee and violated four out of seven commitments made in 2022.

Training Faux Pas

The regulatory body specifically accused Google of failing to negotiate in good faith with publishers based on transparent and non-discriminatory criteria. It also alleged that Google did not provide publishers with the necessary information to assess appropriate compensation for using their content in AI training. Additionally, the authority claimed that Google failed to take measures to ensure that negotiations with publishers did not affect other business relationships between the tech giant and press agencies or publishers.

One key issue highlighted was Google’s alleged practice of tying the display of content on its services, such as the search engine, to whether or not the content was protected from being used in AI training models. The authority stated that Google did not offer publishers an option to opt out of this practice. The fine comes after France imposed a $593 million penalty on Google in 2021 for allegedly disregarding an order to work fairly with publishers regarding the display of news content.

“The Autorité notes that Google failed to comply with its commitment to cooperate with the monitoring trustee, in particular by failing to share all the information necessary for the monitoring trustee to monitor the commitments. The Autorité also notes that Google tried to delay the timing of when the Autorité would be informed in the event that the monitoring trustee had any doubt about a potential breach,” the agency wrote in announcing the decision. “The Autorité found during its investigation that Google used content from the domains of press publishers and news agencies when training the foundation model of its artificial intelligence service and for the grounding (the sending of a query by the artificial intelligence service to Google Search, in order to provide an answer to the question posed by the user) and display (the displaying of the answer to the user) stages, without either the press agencies and publishers or the Autorité being informed of these uses.”

Euro Regulation

France, Germany, and Italy previously all agreed on rules regulating generative AI and large language models (LLMs). The rules included including mandatory self-regulation for generative AI foundation model developers of AI through ethical codes of conduct. Meanwhile, the European Union just recently passed a major new set of regulations surrounding AI, with a focus on generative AI models and their deployment and training. The rules and the new fine underscores Europe’s efforts to set norms for emerging technologies while supporting innovation. Italy’s inclusion is especially notable after the country outright banned ChatGPT for a while over concerns about privacy and intellectual property rights.

France, Germany and Italy Reach AI Regulation Agreement

Italy Lays Out Demands and Deadline for Lifting ChatGPT Ban as EU Opens Investigation

Leading AI Developers Agree to White House’s AI Safety Principles