AI Writing Assistants Are Now Used by 60% of Knowledge Workers
A new survey reveals the majority of office workers use some form of AI writing tool weekly, up from 28% just two years ago.
A new survey of 4,200 knowledge workers across industries finds that 60% now use AI writing tools on at least a weekly basis, more than doubling adoption rates from two years ago. The data suggests AI writing assistance has crossed from early-adopter novelty into mainstream workplace practice — a shift that carries implications for how organizations think about productivity, skill development, and content quality.
What workers are actually using AI for
The survey breaks down use cases in ways that complicate the popular narrative of AI replacing writers. The most common applications are editing and polishing existing drafts (73% of AI writing users), generating first-draft outlines (61%), and rephrasing sentences for clarity or tone (58%). Fully AI-generated content that goes out without human review is reported by only 18% of respondents — and predominantly for low-stakes internal communications.
The email effect
Email is far and away the most common AI-assisted writing format, cited by 81% of AI tool users. Respondents describe using AI primarily to shorten lengthy emails, adjust tone for sensitive conversations, and draft replies to complex requests when they're short on time. Several respondents noted that AI assistance has made them more responsive — not by replacing their judgment, but by reducing the friction of getting words on the screen.
Concerns about quality and authenticity
The survey also captures significant ambivalence. 44% of respondents who use AI writing tools report concerns that their writing has become "less distinctive" since they started relying on AI assistance. Managers surveyed separately express concern about homogenization — a fear that AI-polished communications are starting to sound similar regardless of the author. Several organizations report implementing "AI disclosure" norms for certain document types.
Implications for writing tools
For standalone word count and writing utilities like WordTally, the adoption data is relevant context. Writers using AI assistance still need to track, measure, and edit their output — often more carefully than before, since AI-generated text can be verbose or imprecise in ways that manual writing rarely is. The growth in AI writing use has, counterintuitively, increased demand for editing and measurement tools.