AI plugin adds typos to emails

A new Chrome plugin called Sinceerly promises to make AI-written emails look less like AI by adding mistakes on purpose. According to Business Insider, the tool rewrites polished text to introduce typos and awkwardness, offering modes labelled “Subtle”, “Human”, and “CEO” for progressively shorter and sloppier output.

The product sits on top of a now-familiar workflow: workers draft a message with a large language model, then try to sand down the telltale signs that it was machine-made. In the first wave of generative writing, the giveaway was overlong politeness and synthetic rhythm; in the second, it is the absence of friction—no misspellings, no odd commas, no half-finished thought. If the medium has taught recipients to treat flawless prose as suspicious, then a typo becomes a credibility signal, even when it is generated.

That shift matters because it erodes one of the few low-cost ways humans have had to infer effort and authorship in digital communication. A typo used to be a minor defect; now it can function as a kind of watermark for “this was typed by a person under time pressure”. Tools like Sinceerly turn that watermark into a commodity, available to anyone willing to pay for a browser extension.

The same logic is spreading beyond email. Platforms, employers and schools have leaned on “AI detection” as a backstop for authenticity, but the market has moved toward evasion: rewriting, paraphrasing, and now deliberate imperfection. When the incentive is to look human rather than to be human, the easiest path is to imitate the surface features that readers associate with human work.

Sinceerly’s pitch is small, even comic, but it points to a larger pattern: AI systems are not just generating content; they are generating the social cues that content is supposed to carry. In the inbox, that cue is a typo.

The plugin’s value proposition is simple: it sells plausible messiness to people who already outsourced the first draft.