Artist Platform Scraped by Data Miner; Former Scraper Joins Effort to Build Defense

Cara, a portfolio and image-sharing service founded by photographer Jingna Zhang and a volunteer team, attracts roughly 1.5 million visual artists who seek protection from unauthorized AI training. The platform filters AI-generated content and provides features like Glaze to mask artistic style, yet it cannot fully prevent data harvesting. In August, three distinct scrapers extracted large portions of the site’s publicly available library, posting a 12-terabyte archive on a Reddit forum, uploading link collections to Hugging Face, and distributing images and user data via Academic Torrents.

These breaches spiked Cara’s hosting expenses and prompted Zhang to start a GoFundMe for legal costs, quickly raising over $100,000. Confronting the first scraper, who later apologized, led to a partnership in which the former miner helps identify vulnerabilities and develop an open-source alert system that notifies artists when their work surfaces in AI datasets. Zhang hopes the episode will draw policy attention to the broader problem of web-wide scraping and AI misuse.