CJEU: First referral on copyright vs. AI

Key questions before the CJEU are as follows:

1. Does it constitute a "communication to the public" when an LLM-based chatbot displays text that partially matches content from newspaper publishers' websites? 
   
   If question 1 is answered affirmatively, does it matter that this is the result of a process where the chatbot merely predicts the next word based on observed patterns?
    
2. Does the learning process for an LLM-based chatbot constitute reproduction when this process is built on observation and pattern matching?
   
   If question 2 is answered affirmatively, is such reproduction covered by the exception in Article 4 of the DSM-directive (Directive 2019/790) regarding the copyright exception for text and data mining purposes?
    
3. Does it constitute a reproduction when a chatbot generates a response based on user instructions that reference text in a press publication?

The court's ruling could set precedents for how AI systems can legally interact with copyrighted content, particularly news and press materials. The outcome will likely influence future AI development practices and could reshape the relationship between technology companies and content publishers in the digital sphere. Schjødt is closely monitoring this case as it may establish crucial boundaries for AI training and content generation across the European Union.