TLDR
Anthropic's updated Claude usage policy, effective 12 November 2026, makes it a direct breach to use Claude to seed fake sites, personas or reviews that manipulate what search engines and AI systems cite. Its September threat report described a French ad agency's network of about 70 fake news sites built to lift search rankings. Australian agencies selling Claude-powered content networks or fake review programs are now in breach of the terms.
KEY TAKEAWAYS
Anthropic has banned the use of Claude to "manipulate the sources from which search engines or AI systems draw answers", a rule aimed squarely at the fake-site networks some agencies sell as AI visibility services.[1]
The clause sits in an updated usage policy Anthropic published on 8 October. It takes effect on 12 November.[2]
It covers seeding search engines and AI systems with content that "misrepresents its origin, authorship, or independence", Search Engine Journal reported. The example given is a network of sites posing as unaffiliated sources that back up the same claims.[1]
Fake personas, outlets and reviews are also barred, along with selling those tactics to clients as a service.[1]
A French agency and 70 fake news sites
Anthropic had already acted on one such network. Its September threat report described about 70 fabricated news websites run by LKM Company, a France-based digital advertising agency.[3]
"The articles were specifically designed to boost their site's authority rankings on search engines," the report said. Anthropic said it banned the accounts and "the organization associated with this activity".[3]
The new rules sit in a section titled "Do Not Engage in Deceptive Campaigns or Artificial Activity", which applies to deception "of any kind (whether political or commercial)".[2]
"Most of the updates in the latest version are intended to clarify existing rules," Anthropic said.[2]
Not every change tightens things. Automated content publishing is no longer on the policy's high-risk list, though SEJ noted that is "a change to the high-risk list, not a blanket exemption from the policy". Anthropic also removed its blanket ban on personalised vote and campaign targeting.[1][2]
Agencies, not websites, carry the risk
Google's spam policies act against websites. Anthropic's rule acts against Claude users, SEJ pointed out, so an agency that builds a fake-source network with Claude risks its account whether or not the sites ever rank.[1]
That applies to Australian SEO and generative engine optimisation (GEO) agencies as much as anyone. The test in the policy is deception about where content came from or who is behind it, so ordinary content marketing and digital PR under a client's own name aren't caught.[4]
Agencies running review programs or networks of "independent" sites on Claude have until 12 November to change them.[2]
SOURCES & CITATIONS
FREQUENTLY ASKED QUESTIONS
What exactly does Anthropic's new policy ban?
Does the policy affect standard content marketing or SEO?
When does the updated policy take effect?
Has Anthropic already taken action against fake source networks?
Xaviery Malinao writes for Prompt the Market on how brands and agencies are adapting to answer engines, drawing on Bushnote's work with clients across search, AI search and content.
