Wikipedia volunteers spent years cataloging AI tells. Now there's a plugin to avoid them.

"To work around those rules, the Humanizer skill tells Claude to replace inflated language with plain facts and offers this example transformation: Before: "The Statistical Institute of Catalonia was officially established in 1989, marking a pivotal moment in the evolution of regional statistics in Spain." After: "The Statistical Institute of Catalonia was established in 1989 to collect and publish regional statistics." Claude will read that and do its best as a pattern-matching machine to create an output that matches the context of the conversation or task at hand."

"Even with such a confident set of rules crafted by Wikipedia editors, we've previously written about why AI writing detectors don't work reliably: There is nothing inherently unique about human writing that reliably differentiates it from LLM writing. One reason is that even though most AI language models tend toward certain types of language, they can also be prompted to avoid them, as with the Humanizer skill. (Although sometimes it's very difficult, as OpenAI found in its yearslong struggle against the em dash.)"

A Humanizer skill instructs Claude to replace inflated language with plain factual sentences, using example before-and-after transformations to shorten and clarify statements. AI-detection methods are unreliable because no stylistic feature uniquely distinguishes human prose from large language model output. Large models can be prompted to avoid common model-specific phrasing, and humans can produce chatbot-like text, producing overlap. Professional punctuation and conventions, such as em dashes, complicate detection because models learned them from high-quality sources. Guides that list detector tells present observations rather than strict rules, and user studies show high detection rates still produce meaningful false-positive rates that risk discarding valid human-produced content.

Read at Ars Technica