A Harvard Business School student has created a new tool that (ironically) uses AI to make AI-generated emails look more human. Called “Sinceerly,” the Chrome extension and web app add small mistakes and more casual phrasing to messages. That’s what it’s meant to do, anyway; I couldn’t get the Chrome extension to work, even after several troubleshooting steps.
Sinceerly’s creator, Ben Horwitz, describes the tool as the opposite of Grammarly, which fixes grammar and polishes tone across virtually all websites. Since Sinceerly changes text to look a bit rough and hurried, a reader may be more likely to believe a real person wrote it. The service uses an AI model and rules to add typos, trim overly formal phrasing, and vary sentence length, Gizmodo reports.
The tool offers 3 models: Subtle, Human, and CEO. Subtle makes light edits and keeps most of the original structure; Human writes in a more casual style with some visible errors; and CEO sends very brief, straightforward messages, with the brevity of most executive emails.
Some generative AI users already add typos to chatbot-written messages, school materials, or professional outputs to evade detection. Sinceerly just highlights the ongoing back‑and‑forth between detection systems and techniques aimed at hiding AI’s role in writing.
