Patreon, the popular membership platform for digital creators, has significantly strengthened its defences against artificial intelligence (AI) scraping. The company announced on Thursday that it is now actively blocking AI bots designed to train models on creators' work without consent, a shift from its previous reliance on standard 'robots.txt' files that merely ask bots not to scrape.
This enhanced protection comes through an expanded partnership with internet infrastructure giant Cloudflare, leveraging its AI Crawl Control technology. Patreon stated that the more robust measures were necessary due to the growing sophistication of AI scraping techniques, which have evolved considerably since the platform first introduced deterrents in 2023. While Patreon's paywall has historically shielded much content from crawlers, new discovery features like a redesigned Home Feed and its 'Quips' format could potentially expose more material to unauthorised AI training.
The move reflects a broader industry challenge as publishers and content creators grapple with AI models ingesting their work to improve their intelligence. Cloudflare has been at the forefront of offering tools to combat this, including its 'Pay Per Crawl' marketplace which allows websites to charge AI bots for access. Earlier this month, Cloudflare also updated its policies to default block 'mixed-use' crawlers – those that both index and train – on ad-hosting pages.
Drew Rowny, Patreon's product chief, emphasised the company's vision for creator control. He remarked, "As AI agents become increasingly powerful and popular, creators deserve a meaningful say in how their work is used by AI companies." Rowny highlighted that Patreon aims to provide creators with the ability to grow their audience while maintaining control over their content, contrasting with the general internet where creators often have to accept AI training as a condition of reach.
Testing of the new blocking features reportedly saw individual AI training crawlers' weekly attempts to access Patreon drop from thousands to zero, indicating that many bots were previously ignoring the robots.txt directives. Patreon clarified that it will still permit bots that index pages and organise information to direct users back to the platform, differentiating between beneficial indexing and unauthorised AI training.