Cloudflare now shows site owners which AI assistants visit their pages and lets them allow, limit or block each one, so shoppers using agents are not turned away by mistake.
Cloudflare, a company that helps run and protect websites, has added tools that show which personal AI agents visit a site and what they do there. Site owners can also now choose, agent by agent, whether to let them in. The aim is to stop owners from treating all automated visits the same way.
A personal agent is an AI helper that acts for a person online. A customer might ask one to compare products, find support articles or complete a purchase. To a website, all of that looks like automated traffic. Cloudflare says blocking every agent can feel safe, but it can mean losing customers.
In Cloudflare's Security Analytics, a new Personal Assistants category now sits beside labels such as Verified Bot and Likely Human. A new widget lists which agents visit and how many requests each sends. Owners can filter by one agent to see the pages it asks for, the errors it hits and the networks it comes from.
Cloudflare introduced AI bot policies this summer for Search, Training and Agent behavior. The Agents policy now covers individual personal agents, and the setting is renamed AI Crawler and Agent access policies. Each supported agent can be set to Allow (do not block), Block on pages with ads, or Block. Existing Search and Training policies stay as they are.
An owner could set the overall Agents policy to Block on pages with ads, then allow one agent they want to welcome. New sites added to Cloudflare are asked whether they run ads, and the recommended policies follow that answer.
Account Abuse Protection, which is in early access, shows which user each agent acts for. Login and sign-up events show whether an agent made the request, and which one. If an attacker seems to be using an agent on an account, analysts can act on that account with a rule tied to its Hashed User ID, without blocking the agent for everyone else.
Cloudflare says its next goal is helping sites keep context about their users when an agent stands in the middle, such as whether someone is a customer or has a subscription. That will need privacy-aware ways to carry user context, judge reputation and assign accountability. No timeline was given.
If you run a website, check which AI agents already visit, then set your rules in Security Settings under AI Crawler and Agent access policies.
World models, AI that learns physics from video, are being built into home robots and driverless cars, and your own videos may help train them.
Read more: AI that learns how the physical world works is coming to robots and carsA new Alibaba-led test of real coding requests found the best AI agent fully finished just 26.4 percent of tasks, so human review still matters.
Read more: AI coding helpers fully finish only about 1 in 4 real tasksA University of Chicago law professor asked Claude to fix page numbers. It deleted his syllabus and a year of notes instead. Dropbox saved him, not the AI.
Read more: An AI assistant deleted a law professor's year of lecture notes
