People also ask
In practice
Cloaking has always been the clearest line in search: serving one thing to humans and another to the crawler is deception by definition, and the guideline violations that end careers in this industry usually start here. The AI-era version swaps in a new crawler — serving chatbot user-agents extra paragraphs, inflated claims or competitor-diminishing text invisible to readers — and inherits all the old problems plus some new ones.
Detection is easier than practitioners assume: engines compare what the crawler saw with what users and other systems see, run their own verification crawls from different identities, and accept reports; a discrepancy is trivially provable by diffing two fetches. The consequences compound: exclusion from the engine that caught it, devaluation of the whole domain’s content in its answers, and reputational damage in a market where the same editors and buyers watch everything.
The honest framing for a vendor: this is R&D-lab territory at most — understanding that chat crawlers are currently under-policed is a research observation, not a client opportunity — and any agency proposing it for a client is transferring a business-critical risk to a marketing experiment. The legitimate version of the underlying goal already exists: make the real content maximally extractable — clear answers in prose, structured data, consistent entities — so that what the crawler sees and what the reader sees are the same excellent page.
Nothing that requires deceiving the reader about what the crawler reads survives contact with a serious client’s legal review, and nothing that survives it needs the deception.
See also: AI crawler access, Link schemes, Algorithmic demotion.
Related service: International links.
