Patreon blocks AI training bots with Cloudflare, ends polite requests

Patreon's shift from soft deterrence to hard technical blocking reflects a wider industry reckoning with AI model training. When the company first introduced blocking measures in 2023, it relied on robots.txt—a cooperative standard that assumes good faith. But testing revealed that AI training crawlers were simply disregarding those requests and scraping anyway, making the polite approach ineffective. The timing of this upgrade is also driven by changes to Patreon's own product: the newly redesigned Home Feed and the introduction of Quips (short-form content) both threaten to surface more creator work to scrapers, making tighter controls necessary.

This move sits within a broader ecosystem shift. Cloudflare's recent policy change to block "mixed-use" crawlers by default on ad-supported pages, and its introduction of Pay Per Crawl, signal that the infrastructure layer is now actively helping publishers and creators negotiate—or refuse—the terms of AI training access. Patreon's product chief Drew Rowny frames the company's stance around creator agency: most creators online must accept AI training as the cost of reaching an audience, but Patreon's model allows creators to retain control. This positioning suggests that platforms protecting creator interests from uncompensated AI training may become a competitive differentiator.