Humanizer Plugin Uses Wikipedia AI Detection Guide to Alter Claude Output
At a glance
- WikiProject AI Cleanup published a list of 24 AI writing patterns in August 2025
- Over 500 Wikipedia articles have been tagged for possible AI-generated content
- Humanizer plugin for Claude Code uses the Wikipedia guide to adjust AI responses
WikiProject AI Cleanup, a volunteer group of Wikipedia editors, published a formal guide in August 2025 identifying 24 common language and formatting patterns that indicate AI-generated writing. This development led to new tools and processes for detecting and addressing AI-written content on Wikipedia and beyond.
The group has applied their guide by tagging more than 500 Wikipedia articles for review as potentially AI-generated. These actions are part of ongoing efforts to maintain content standards and transparency on the platform.
In response to the publication of the guide, tech entrepreneur Siqi Chen released an open-source plugin called Humanizer for Anthropic’s Claude Code AI assistant. The plugin incorporates the 24 patterns from the Wikipedia guide, instructing the AI to avoid writing in ways that resemble AI-generated text.
The Humanizer plugin is implemented as a Markdown-formatted skill file. This file appends structured instructions to Claude Code prompts, which the system interprets with greater precision than standard prompts. The plugin specifically directs Claude to avoid certain types of language, such as promotional terms, vague attributions, and collaborative phrases like “I hope this helps!”—all of which are flagged in the Wikipedia guide.
What the numbers show
- WikiProject AI Cleanup published its list of 24 patterns in August 2025
- Over 500 Wikipedia articles have been tagged for review as potentially AI-generated
- The Humanizer plugin reached more than 1,600 stars on GitHub shortly after release
According to published information, the Humanizer plugin became available on GitHub and quickly attracted attention from users, accumulating over 1,600 stars by the Monday following its release. This indicates a notable level of interest in tools that help AI systems produce more human-like text.
The plugin is designed to update automatically when the Wikipedia AI-spotting guide is revised. This feature aims to keep the plugin aligned with the latest community standards for identifying AI-generated writing on Wikipedia.
By directly referencing the Wikipedia list, the Humanizer plugin provides a procedural method for Claude Code to reduce the likelihood of generating text that matches known AI patterns. This approach reflects a technical response to the growing use of AI in online content creation.
WikiProject AI Cleanup continues to monitor Wikipedia articles and update its guidance as new patterns of AI-generated writing are identified. The ongoing collaboration between community editors and developers highlights the evolving landscape of content moderation and AI tool development.
* This article is based on publicly available information at the time of writing.
Sources and further reading
More on Technology
-
AI Swarms Raise Concerns Over Disinformation and Democracy
Research in Science highlights AI swarms' involvement in disinformation campaigns. Experts call for an independent AI Influence Observatory.
-
Equity and PACT Continue Talks Over AI Protections for Performers
Equity's ballot showed over 99% of 7,000 performers oppose digital scanning. PACT has since proposed improved AI protections, with talks ongoing.
-
China Expands Role in Global Humanoid Robotics Development
China's robotics industry is booming, with over 200 firms developing humanoid robots and accounting for two-thirds of global patents in 2024.
-
Transformer Architecture Shapes Foundation of Modern AI Models
Introduced in 2017, the Transformer architecture utilizes attention mechanisms. It powers models like BERT and GPT, according to researchers.
-
Humans& Startup Reflects Rise of Elite AI Neolabs
Humans& aims to raise $1 billion at a $5 billion valuation, developing AI models for user collaboration, founded by top lab researchers.