Medium: using AI for content creation without losing your voice
What the article is about
Saketh’s Medium piece addresses a common anxiety among writers and content creators: that using AI tools requires trading authenticity for speed. The article argues this is a false choice. Through a five-step framework and supporting examples, it describes how to position AI as a production tool rather than a voice substitute, maintaining the editorial judgment and personal perspective that distinguish effective content from generic output.
Context
The article is written from a practitioner’s perspective rather than an industry overview, and it is most useful for individual writers, bloggers, and content creators who already understand that pure AI generation produces detectable, flat output. The author’s framework assumes that the creator has something specific to say — their own experiences, data, or relationships — and treats AI as a means of saying it faster and with better structure.
Key takeaway
The five-step framework is organized around the sequence of a writing workflow rather than tool categories. Research and ideation come first: AI is used to gather information, explore angles, and surface questions the creator had not considered, while the final position remains entirely the creator’s. The second step uses AI to overcome blank-page friction by generating a rough first draft, with the understanding that this draft will be substantially rewritten in the creator’s voice before it is published.
Technical optimization comes third: applying AI to SEO, readability scoring, and grammar corrections, which frees the creator to focus on where AI cannot substitute — emotional resonance and specific argument. The fourth step involves adding what AI cannot contribute: personal anecdotes, original data points, first-person observations, and references to specific conversations or interviews. These additions are what give a piece E-E-A-T (experience, expertise, authority, trustworthiness), and they cannot be generated from a prompt.
The fifth step addresses multimedia: combining original photographs, graphics, or audio with AI-assisted text, so that the overall package has elements that are distinctly the creator’s work rather than generated output in visual packaging.
The article includes the example of a finance blogger named Sophia Chen who moved from two to six weekly articles using this approach while reporting a 34% increase in audience engagement — a data point that frames the framework as practical rather than theoretical, though it comes without a link to verify the claim.
Who should read this
Writers, bloggers, and content creators who are already using AI tools but finding that published output lacks the response they were getting from fully human-written work. Also relevant for content managers evaluating frameworks for teams using AI in production workflows.