Brief
Anthropic bans seeding Claude with fake sources to curb deceptive AI answers
The updated policy adds a dedicated clause against manipulating sources that feed AI answers, with possible suspension for violators.
By Felo News Desk · Published
Anthropic announced on Oct. 8 that, starting Nov. 12, its Claude model may not be used to seed search‑engine or AI‑answer sources with content that misrepresents origin, authorship, or independence, according to a post on the company’s site.
What happened
The updated usage policy introduces a new section titled “Do Not Engage in Deceptive Campaigns or Artificial Activity.” It explicitly bars the manipulation of sources that AI systems draw answers from by planting fabricated articles, networks of sites posing as independent outlets, or fake reviews. The rule also forbids creating tools for such deceptive campaigns or selling them as a service. Violations could lead to suspension or termination of access for developers using the API or users of Claude‑powered apps.
What the report adds
SearchEngineJournal notes that the policy now covers both political and commercial deception, whereas the earlier version (effective Sept. 15, 2025) did not mention search engines or AI responses. The post lists two specific practices: seeding AI answer sources with misleading content and operating networks of sites that hide a common origin. It also reiterates that fake personas, fake news sites, and fake reviews remain prohibited, as they were in prior rules.
What was said
The company’s announcement states, “Manipulate the sources from which search engines or AI systems draw answers by seeding them with content that misrepresents its origin, authorship, or independence (e.g., networks of sites posing as unaffiliated sources corroborating the same claims).” The wording appears in the policy text and is quoted in the SearchEngineJournal article.
How it came about
Anthropic’s earlier policy updates in 2025 addressed elections, fraud, privacy and disinformation but left the search‑engine clause vague. The Oct. 8 post consolidates those scattered provisions into a single section, reflecting growing concerns about AI‑generated misinformation. This follows Anthropic’s recent product announcements, such as the integration of Claude Chat and Cowork and the rollout of Claude Code mods, which have expanded the model’s capabilities across developer and knowledge‑worker workflows.
Key facts
- Anthropic's updated usage policy takes effect on Nov. 12, 2026. (searchenginejournal.com)
- The new policy adds a section titled “Do Not Engage in Deceptive Campaigns or Artificial Activity.” (searchenginejournal.com)
- Violations may result in suspension or termination of Claude access. (searchenginejournal.com)
- The policy now explicitly bans seeding AI answer sources with fabricated content and operating networks of fake independent sites. (searchenginejournal.com)
- Earlier version (Sept. 15, 2025) did not reference search engines or AI responses. (searchenginejournal.com)
Sources
- [1] searchenginejournal.com — originally reported as “Claude's New Rules Target Fake Sources Built To Sway AI Answers”
