3 ms·
>is evidence that its recreating images from its training set No it isn't lol. a shutterstock logo is just more common ground for the model because of how ofte
by famouswaffles 3y ago
>is evidence that its recreating images from its training set
No it isn't lol. a shutterstock logo is just more common ground for the model because of how often it will appear in the dataset.
- p1necone 3y agoYou don't see how it generating a registered trademark in its output is a problem? You think it's okay to monkey patch this one obvious case and call it good?
- andromeduck 3y agoIf you think it's the trademark that's the issue then removing it is sufficient.
- benterix 3y agoI think you misinterpreted parent's statement.
- madeofpalk 3y agoThey do not think it's only that one trademark that's the issue.
- michaelt 3y agoShutterstock would say that the model is a derivative work of their images, much like a photograph of a painting is a derivative work of the painting despite being in a different medium. They would say they are owed licensing fees regardless of whether the shutterstock logo appears in the output. The appearance of their logo in the output merely proves that the system's outputs are derived from their copyrighted images.
- singleshot_ 3y agoWouldn’t the appearance of shutterstock’s mark suggest to the consumer that the provenance of the image was shutterstock, when it is not? But upon the absence of the mark, shutterstock would argue that the provenance of the image WAS shutterstock? I don’t think this is the salient feature of this phenomenon, but the proximity of these two arguments is interesting.