Blocking ChatGPT and OpenAI crawlers: what it means for your traffic

WebRankInfo published a guide on how to (try to) block OpenAI and ChatGPT from using a site's content, noting that OpenAI's tools can draw on site content to learn and to supply answers. That is the whole of what the report states: the intent is possible, the blocking is an attempt rather than a guarantee. The publication does not specify which mechanisms it recommends, which user agents or controls are involved, or what level of effectiveness can be expected.

Why the decision is now an acquisition question

For years, blocking a crawler was a narrow technical choice, mostly about bandwidth or duplicate content. AI answers change the calculus. If a model summarises or answers from your pages, a reader may get the substance without visiting the site, which affects sessions, ad impressions and conversion opportunities. Blocking is the lever a publisher controls directly, and it is one of the few available before any commercial arrangement exists.

Caveat: the guidance is described as an attempt. Treat any block as risk reduction, not as enforcement. Crawlers change, access paths multiply, and content can reach models through other routes. The report does not quantify how much traffic is at stake, and neither should any internal forecast built on it without your own data.

What to check this week

Start with measurement, not with a block. A short audit is enough to decide.

• Pull server logs and analytics referrers for the last 90 days to see whether AI-associated traffic is measurable at all, and how it converts compared with search and social.
• List the page types where being quoted without a click is acceptable (definitions, glossaries) and those where it is not (pricing, proprietary data, gated assets).
• Check whether your robots directives and any AI-specific controls are actually deployed in production, not only in staging.
• Record a baseline before changing anything: organic sessions, assisted conversions, and revenue per session by page type.

The monetisation side

Blocking is one strategy among several. Others include licensing content, keeping high-value material behind authentication, or accepting AI citation as a brand channel while monetising elsewhere in the funnel. None of these is settled practice, and the WebRankInfo report does not evaluate them. What it does establish is that the choice is now explicit: a site owner can attempt to exclude OpenAI's use of their content, and should decide deliberately rather than by default.

Whichever route you take, document the decision and its date. When the crawler landscape shifts again, a dated baseline is the only way to tell whether the call was right.