AI Tool Intentionally Adds Typos to Emails for Human Authenticity

Large language models, the underlying technology for advanced AI writing assistants, now produce highly polished writing. This rise in AI-generated text has created suspicion around overly tidy digital communications. Consequently, a new approach seeks to reintroduce human imperfections into AI-assisted writing.

Venture capitalist Ben Horwitz developed an "anti-Grammarly" tool called Sinceerly. This browser plugin uses AI to deliberately add typos and other errors to emails. Horwitz stated the tool's purpose is to "mess up your emails with AI," counteracting the pervasive perfection of AI-driven grammar checkers. Sinceerly can alter capitalization, introduce misspellings, and append phrases like "sent from my iPhone." Its settings include a "CEO" mode, designed to mimic the typically terse and error-prone communications of high-ranking executives. Horwitz reported testing Sinceerly by emailing five Fortune 500 CEOs. Four of them responded, with each reply containing fewer than ten words. Notably, two of these CEO replies also featured typos, suggesting an unintended mirroring effect. However, the Sinceerly plugin developed bugs after its initial public exposure, rendering it largely unusable.

This tool highlights a growing tension between efficiency and authenticity in digital communication. As AI refines writing, the very imperfections it removes become markers of human origin. The attempt to use AI to simulate human flaw suggests a new frontier in online interaction. While Sinceerly currently faces technical issues, its concept indicates a developing trend. Future developments may refine AI's ability to create intentionally imperfect, yet authentically perceived, digital text.