Anthropic's Usage Policy update that bans using Claude to seed fake or disguised sources designed to manipulate what search engines and AI systems report — part of a new "Do Not Engage in Deceptive Campaigns or Artificial Activity" section.
Last verified · 2026-10-10 · by Moe Ameen
"Claude's fake-source rules" is shorthand for a section Anthropic added to its Usage Policy in an update published October 8, 2026, with an effective date of November 12, 2026. The section is titled "Do Not Engage in Deceptive Campaigns or Artificial Activity," and the clause most relevant to content people prohibits using Claude to manipulate the sources that search engines or AI systems draw answers from by seeding them with content that misrepresents its origin, authorship, or independence.
The rule is wider than that one sentence. It also bans creating fake personas or media outlets to mislead audiences, distributing content through networks of fake or ostensibly independent outlets, websites, or accounts so their shared source stays hidden, and building tools for — or selling as a service — this kind of coordinated inauthentic behavior. The previous Usage Policy, last updated in 2025, did not mention search engines or AI answers; this is a net-new line. It binds everyone using Anthropic's products, including API developers and apps built on Claude, and it is enforced with warnings, usage limits, suspension, and account termination.
It helps to know what prompted it. Anthropic's threat-intelligence reporting described a France-linked operation that ran roughly 70 websites posing as independent local-news outlets and used Claude to mass-produce articles in a fixed template — each carrying three to four internal links — specifically to inflate the sites' authority signals with search engines. Anthropic says at least 8,913 articles were published across about 20 languages, that most drew little real engagement, and that it removed the account and banned the organization. The new policy line generalizes that takedown into a standing rule.
The honest framing for a creator: this is not a limit on what Claude can write or a watermark on its output. It is a rule about how you may use that output. If you publish AI-assisted content openly under your own brand, nothing here applies. If you were planning to farm AI citations with a ring of look-alike "independent" sites, that plan is now a policy violation. Treat the threat-report figures as Anthropic's own numbers and read the Usage Policy text directly if exact wording matters.
The useful way to read this rule is as a map of the dead end, which makes the honest road obvious. The banned pattern is specific: 70 disposable sites pretending to be unaffiliated, quietly linking to each other to manufacture authority a real brand hasn't earned. The sanctioned pattern is its mirror image — one real identity, publishing attributable content, consistently, where people and AI systems can see exactly who is behind it. [Kompozy](/) is built to run that second pattern at the same scale the fake-network operators were chasing, minus the deception. A [Persona Brief](/glossary/persona-brief) encodes one transparent brand voice, and every piece ships under that same identity: there is no second persona to disguise, because the whole point is that it is unmistakably you.
Where the banned playbook needed volume across many hidden properties, Kompozy gets volume from one owned presence. Feed it a single source and it generates a week of content — avatar [Persona Shorts](/glossary/persona-shorts), brand-exact [Carousel Posts](/glossary/hyperframes), quote graphics, native text posts, a blog article on your own domain, and an email newsletter — then schedules and publishes it across the eight social platforms plus blog and email behind a per-post review gate. That is scale that compounds into real authority instead of evaporating the moment a trust-and-safety team notices the pattern. Keep Claude for the thinking and drafting it is genuinely good at; let an engine designed for open, on-brand distribution be the thing that actually ships it.
In a Usage Policy update published October 8, 2026 (effective November 12, 2026), Anthropic added a section titled "Do Not Engage in Deceptive Campaigns or Artificial Activity." It bans using Claude to manipulate the sources search engines or AI systems draw answers from by seeding them with content that misrepresents its origin, authorship, or independence, and it prohibits fake personas, fake outlets, networks of ostensibly independent sites, and selling such campaigns as a service.
No. They restrict deception, not AI writing. You can still use Claude to draft and research content you publish openly under your own brand. What is prohibited is disguising who made the content or building a network of fake "independent" sources to inflate search or AI-answer authority.
It applies to everyone using Anthropic's products — including API developers and apps built on Claude — not to search engines. Anthropic enforces it with warnings, usage limits, suspension, and account termination if it finds a violation.
Build one real, attributable presence instead of a disguised network. An engine like Kompozy generates on-brand video, carousels, images, a blog, and a newsletter from a single source under one transparent Persona Brief voice, then publishes across the eight social platforms plus blog and email — authority you own rather than fake.