Users of Midjourney text-to-image site claim issues with new update

Text-to-image generator Midjourney began alpha-testing its newest version-6 models on Dec. 21. Users quickly claimed that the update, which seemed to scale up the generator's capabilities, includes results that seem too similar to known copyrighted art. Reid Southen, a concept artist and illustrator who has worked with some of the biggest film studios, including Marvel and DC, said Friday that the model's output for certain prompts is an unaltered iteration of copyrighted training data.