AI has been a hot topic in recent years, with major internet companies beginning to invest in and research their own AI systems. A crucial step for AI development is to "train" the AI using a vast amount of material, which is indispensable. However, this process raises several issues, such as whether using publicly available material for training AI involves copyright infringement.
For Google, AI is an unmissable field of exploration. They started the internal development and training of their own AI as early as 2019. Eventually, under the scrutiny of French authorities, it was found that Google indeed had violated certain laws by using related materials for AI training. Following this discovery, the French authorities demanded that both parties negotiate compensation for the use of the materials.
However, after the intervention of French officials, Google did not take the matter seriously and continued using the same methods for AI training. This behavior once again caught the attention of the French Tax Authority, leading to a ruling in July 2021 where Google was fined 500 million euros and ordered to comply with the initial injunction.
Eventually, the dispute between the two parties came to a temporary halt in 2022, with a settlement being reached. Google made promises, including setting up a framework for negotiation and sharing transparent assessments of the required compensation for rights and making these commitments binding. Essentially, Google needed to allow the licensors access to certain information to determine the compensation they were due and to restrain their own breaches of contract.
However, Google clearly did not take the French Tax Authority seriously, failing to fulfill the promises made. They were recently found to have breached four of the seven commitments they made, including:
- To negotiate sincerely based on transparent, objective, and non-discriminatory criteria within three months (Commitments 1 and 4);
- To provide news organizations and publishers with transparent assessments of the information needed to determine their rights compensation (Commitment 2);
- To take the necessary steps to ensure that negotiations do not affect other economic relationships between Google and news organizations and publishers (Commitment 6).
In 2023, Google officially launched its own AI, "Bard". According to the investigation by the French Tax Authority, Google continued to use the materials of news organizations and publishers for training this AI without notifying the copyright holders or the French Tax Authority, nor did it avoid infringing the copyrights of these organizations. Instead, it used copyrighted content for AI training while also blocking these organizations from claiming compensation, as they could not ascertain how much material was used by Google.
Based on Google's abovementioned breaches, the French side issued another fine of 250 million euros, marking the fourth judgment on this case in four years. The previous three judgments were not taken seriously by Google. Will the fourth ruling prompt them to change their ways?
