Anyone using AI chatbots like Claude who wants to improve the naturalness and quality of their AI-generated text.
Don't fix AI mistakes after Claude writes. The solution is to set up Claude before it generates any text.
'Humanize this' doesn't work because Claude doesn't know your definition of human. It needs a clear filter.
Use ChatGPT to research common AI writing patterns, words, and phrases to create a foundation for your filter.
Organize research into a document listing AI patterns to avoid and human alternatives. This becomes Claude's blacklist.
Convert your organized document to markdown format (anti-ai-writing.md) to upload and use within Claude.
Compare Claude's output before and after uploading the anti-AI writing file. Notice the improved, more natural tone.
Turn your anti-AI writing file into a reusable Claude skill for automatic filtering in all your chats.