Would you be interested in providing training data to make our models more efficient?
For Example: Would you provide training data for Video API to recognize more entities? The new model will be available for all users. This will be a separate end point with separate Terms of Service. Video API or other ML APIs do not use your videos to train Google models.