Yes, you have only to look at the images generated by GANs to see that it's not inherent in the algorithm: images generated by GANs are not just a patchwork of

Yes, you have only to look at the images generated by GANs to see that it's not inherent in the algorithm: images generated by GANs are not just a patchwork of textures, they have coherent shapes. They 'needed' shape in order to fool the discriminator, which in turn developed a sensitivity to shape in order to identify generated images (which probably do rely on texture early on) as fakes.