Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
A new tool aims to help AI chatbots create more human-looking text – with the help of Wikipedia’s guide to AI discovery, As I mentioned Ars Technica. Developer Siqi Chen says he created the tool, called Humanizer, by providing Anthropic’s Claude with a list of stories compiled by volunteer Wikipedia editors as part of an initiative to combat “poorly written content generated by artificial intelligence.”
Wikipedia guide Contains a list of signs that text may be AI-generated, including vague attribution, promotional language such as calling something “amazing,” and collaborative phrases, such as “I hope this helps!” The Humanizer tool, a custom skill for Cloud Code, is supposed to help the AI Assistant avoid detection by removing “AI-generated typing marks from text, making it appear more natural and human.” According to its GitHub page.
The GitHub page provides some examples of how Humanizer can help Claude uncover some of these tidbits, including by changing a sentence that describes a location as “located within a picturesque area” to “a town in the Gondar region,” as well as modifying an ambiguous attribution, such as “experts believe it plays a crucial role” to “according to a 2019 survey by…” The guide is updated.
It’s only a matter of time before AI companies themselves start adjusting their chatbots to some of this news as well, as OpenAI has done We’ve already addressed the overuse of ChatGPT Of em dashes, which have become an indicator of the content of the artificial intelligence.