We have published a slightly longer form article looking at a recent decision on the interplay between ‘text and data mining’ (TDM) exceptions to copyright and the creation of datasets for training AI models, and how rightsholders can opt out of their works being used. The key points are:

• A German court recently considered a case where a photographer sued a non-profit for using his images in an AI training dataset. The defendant, LAION, had created a massive dataset for training AI models, including the claimant photographer’s works. 
• The ruling gives guidance on how EU and German laws handle TDM for AI, and specifically the rights of creators to opt out of their work being used. 
• The court hinted that website terms and conditions might be a valid way for creators to opt out of having their work used in AI training.
• This decision underscores the uncertainty around how rightsholders can opt out of their works being used for training, and suggests some ways that they could do so.