Key points:
Background on GenAI Model Training: Generative AI models, particularly text-to-image models like Midjourney and OpenAI’s DALL-E 3, have the ability to create diverse artworks based on textual descriptions. However, many of these models have been trained on artwork without artists' knowledge or permission, prompting the need for tools to protect artists' rights.
Introduction of Kin.art’s Tool: Kin.art has launched a free tool designed to prevent GenAI models from training on artwork without the artists' consent. This tool employs image segmentation and tag randomization techniques to interfere with the model training process. Unlike some existing solutions, Kin.art's tool focuses on preventing artwork from being included in datasets in the first place, rather than mitigating damage after inclusion.
Philanthropic Approach and Future Plans: Kin.art’s co-founder, Flor Ronsmans De Vry, positions the tool as a philanthropic effort, emphasizing the importance of an ethical approach to AI training that respects artists' rights. While the tool is free, artists are required to upload their artwork to Kin.art’s portfolio platform. The company plans to offer the tool as a service in the future, aiming to help various platforms protect their data from unlicensed use in the age of AI.
Discussion (1)
The success of this strategy hinges on widespread adoption by external platforms, which may be a challenging for any type of organization that does this. I know startups doing generative AI would likely need government intervention for this to be something that actually takes off, I highly doubt they will seek this solution themselves otherwise.