Cloudflare Adds a 'Disallow AI Training' Setting That Keeps You in Search, With Apple, Google and Microsoft Certified 'Accountable'

Cloudflare Blog·high signalXBlueskyLinkedInCopy link

Cloudflare shipped the control on September 15 to break the mixed-use crawler bind, where one bot both indexes for search and scrapes for training. Operators earn 'Accountable' status by meeting four requirements — a real opt-out, page-level transparency into training data, a guarantee that opting out does not hurt search rank, and AI-summary controls — and Apple, Google and Microsoft passed; non-compliant crawlers get blocked while the preference is published in robots.txt. Cloudflare's own number frames the demand: 17% of sites already block training in some form, against under 1% that block search bots.