# Harvard student builds browser tool that adds typos to AI writing to make it seem more human > A browser plugin called Sinceerly deliberately adds typos to AI-generated text so emails read as human. Flawless writing now signals machine authorship, making small errors a trust signal. - Type: AI news - Page: https://completeaitraining.com/news/harvard-student-builds-browser-tool-that-adds-typos-to-ai/ - Published: 2026-04-27 - Updated: 2026-04-27 - Author: Jeroen Erne - Jobs: Writers - Reading time: 2 min ### A Browser Tool Adds Typos on Purpose to Make AI Writing Feel Human A new browser plugin called Sinceerly does something counterintuitive: it makes AI-generated text worse. The tool sprinkles in small typos and quirks to make polished emails feel more authentic in a world where flawless writing increasingly signals a machine wrote it. The name itself demonstrates the concept-it swaps an 'e' for an 'i.' That misspelling is the entire point. #### Why Perfect Writing Now Reads as Suspicious AI writing tools produce grammatically flawless text in seconds. That perfection has become a liability. Readers now associate pristine prose with machine generation, making genuinely human writing feel less trustworthy when it's too clean. AI-generated text often follows predictable rhythms-balanced, measured, never rough around the edges. The polish itself becomes a tell. Sinceerly addresses this by introducing intentional imperfections. A missed letter. A slightly awkward phrase. The kinds of quirks that appear in actual human communication, especially in work emails dashed off between meetings. #### How the Tool Works Sinceerly offers different modes for different contexts. One mode keeps writing mostly polished while adding a mistake or two. Another shifts toward casual tone-the speed of someone typing quickly. A third adopts blunt, short-form style with "sent from my phone" energy. The tool doesn't randomly scatter errors. It mimics patterns of actual human writing, where imperfection is the norm, not the exception. #### The Broader Shift in Professional Communication People have already been manually adding typos to AI-generated content to make it seem more human. Sinceerly simply automates what writers were doing by hand. This reflects a real change in how authenticity reads. Slightly flawed writing now feels more relatable and trustworthy than flawless prose. The goal has shifted from eliminating mistakes to strategically including them. For [writers working with AI](https://completeaitraining.com/tag/writers/), this presents a new consideration: how to balance the efficiency of AI generation with the human touch that builds credibility. Understanding [prompt engineering](https://completeaitraining.com/tag/prompt-engineering/) helps writers get better initial output, but tools like this address a separate problem-the perception of authenticity itself. #### The Irony Is Complete Decades of spellcheck, autocorrect, and grammar tools trained people to eliminate mistakes. Now a tool reintroduces them deliberately. It's a full circle: using AI to write, then using another tool to make that AI writing look less like AI. In a world where machines write nearly perfectly, imperfection has become the marker of something real. --- This is the Markdown form of https://completeaitraining.com/news/harvard-student-builds-browser-tool-that-adds-typos-to-ai/. The whole site is mapped at https://completeaitraining.com/llms.txt.