Cloudflare announced a new Disallow AI Training setting allowing website owners to stay indexed for search while blocking AI training on their content. Apple, Google, and Microsoft honor or have committed to honor this control. The feature addresses the problem of mixed-use crawlers — single bots serving both search and training — which previously forced a binary choice between discoverability and data protection.

Cloudflare introduced an Accountable designation for bot operators meeting requirements: opt-out mechanisms for training and AI summaries, URL-level visibility into crawled pages, and assurance that opting out won't affect search rankings. Apple, Google, and Microsoft qualify. The company reports 17% of sites enable training blocks, while less than 1% block search bots.

New security controls classify bots by behavior: Search, Training, and Agent. Mixed-use crawlers combine Search and Training. Cloudflare identifies, classifies, and blocks non-compliant crawlers, reporting operator behavior on Radar. The goal by early next year is centralized control over AI summary inclusion, set once on Cloudflare rather than per operator.