AI copyright analysis in C-250/25 Like Company v Google may be moving away from a strict input-output divide. The hearing raised a focused issue: whether copyright protection for AI-related acts should be assessed by separating the material fed into a system from the material it produces. That distinction matters because it affects how liability is characterised in copyright disputes involving AI systems.
The legal significance lies in the way such a split can shape the analysis of infringement. If the input and output stages are treated as separate, the legal assessment may focus narrowly on where protected works enter the system and where alleged copying emerges. If that distinction is weakened, the court may be prepared to look more closely at the overall process by which the system uses protected material and generates results. That shift would have practical consequences for how claimants frame their case and how defendants structure their response.
For AI copyright disputes, the point is not merely academic. A rigid input-output approach can make it easier to identify a single act for challenge, but it may also miss the broader mechanics of how the system operates. A more integrated approach could place greater weight on the connection between the use of material at the input stage and the eventual output, particularly where the complaint is that protected works are being used in a way that remains legally relevant across the whole process. That would make the legal test more sensitive to the realities of AI systems, rather than relying on a formal separation that may not reflect how the technology functions.
For UK practitioners, the immediate significance is cautionary rather than definitive, because the case is before the CJEU and no final outcome is identified here. Even so, the hearing suggests that copyright arguments involving AI may increasingly turn on process-based analysis rather than a simple split between what goes in and what comes out, which would increase uncertainty in how infringement claims are pleaded and defended.
The risk is that a move away from the input-output dichotomy could broaden the scope of copyright disputes involving AI and make the legal analysis more demanding at both the evidential and conceptual level.
Disclaimer: This post is for general information only and does not constitute legal advice. Specific advice should be sought for your particular circumstances.
Source: