This tool could “fully shield” artists from AI scraping their artwork
Kin.art introduces a proactive tool that prevents AI systems from scraping artists' work by disrupting the image-label pairing essential for machine learning. Unlike reactive methods that damage outputs, this approach blocks data ingestion entirely, ensuring creations never enter training datasets in the first place. This lightweight solution is computationally efficient, allowing the company to offer it for free compared to expensive alternatives. By focusing on prevention rather than mitigation, it empowers creators to maintain control over their digital assets without the heavy resource costs associated with existing protective measures. This is highly relevant to open data as it challenges the unrestricted availability of creative works for AI training. It highlights the growing tension between the open access of data needed to build models and the ethical necessity of respecting intellectual property rights, urging a reevaluation of how public datasets are curated and utilized.
Source: itsnicethat.comPublished on 2024-01-25