Disney’s use of AI to identify characters, props and locations in animated productions in view of cataloguing and identification make sense. However as one can read the post linked below, if one consider using comparative analysis of faces design to extract features that are not yet ‘exploited’ then this will make possible to block with IP all versions of an existing cartoon character into new shapes and art design all keeping the IP from the original design or ‘seed’.
More this could block in the future all innovative design by creators other than Disney’s because the AI will identify all possible versioning or adaptation to an original design…
Cataloguing is a good thing that the industry needs, but such a technology could create a capacity of appropriation of all possible alternative design for any cartoon character having similar traits…