Shawn Shan builds tools to help artists fight back against exploitative AI
AI summary of the linked articleShawn Shan, a University of Chicago PhD student, builds tools that let artists protect their work from being used to train AI models without consent. He coded the algorithm behind Glaze, which lets artists mask their personal style from AI mimicry and came out in early 2023, and his team introduced Nightshade, which adds invisible "poison" to images that can disrupt models that train on them. Glaze has been downloaded nearly 3.5 million times and Nightshade over 700,000 times. Shan says he wants to shift power from large corporations back to people and plans tools to help regulators audit AI models. |
1 comment mishaSubmitter Go Glaze your art and protect it from style mimicry by text-to-image models and get started with Nightshade to poison text-to-image generative models: https://glaze.cs.uchicago.edu/aboutus.html |
More from technologyreview.com on Kiwi News
|