Cloudflare published a position paper on November 5, written by Bryan Becker, that lists 14 problems AI agents create for websites and concludes that "in nearly every challenge above, the Site bears the primary cost," PPC Land reported. Its examples include a shop that cannot tell whether a purchase was authorized and a publisher whose article is summarized by a bot that never loads an ad. Cloudflare proposes that the user-site relationship and its security duties survive any software in between, favors Web Bot Auth for agent identity and OAuth for accounts, and calls for a machine-readable way to state what agents may do, saying robots.txt was not built for the job.
Cloudflare says three controls are live for customers: analytics for personal assistant traffic, per-agent access policies that can block agents only on pages carrying ads, and Account Abuse Protection in early access. A coalition called Trusted Agent Delegation will take the issues to standards bodies, though no members or dates were named. Cloudflare published no traffic, dispute or adoption figures, and PPC Land noted that the company sells bot management. PPC Land cited Arkose Labs data that 97.7% of legitimate agent sessions carried no verified identity, and HUMAN data that Meta's Muse accounted for 72% of AI agent requests.
Meta added "spaces" to Muse on November 6, self-updating tools that can track prices, Instagram saves and Marketplace messages. Three of Meta's nine examples involve commerce, including a shopping hub and an order tracker for small sellers. On November 5 Meta began rolling out voice conversations and avatars in Muse in the US. Meta has not said whether spaces data will feed ad targeting or whether transactions will carry fees.