In a recent LinkedIn post, Kieran Flanagan discusses why marketers should be cautious about using AI to generate their ideas and words. Flanagan, who says he previously taught people to build AI writing workflows, describes a change in his own view: as the tools and systems grew more sophisticated, he felt creativity had become an engineering problem.
“I’d turned creativity into an engineering problem.”
Flanagan’s case for creative practice
Flanagan argues that original content comes from developing ideas through lived experience, reflection and conversation—not simply from asking a model for polished suggestions. In his view, relying on AI at the earliest stage can make it harder for writers to begin without a tool and can steer their work toward familiar, broadly suitable answers.
He contrasts those outputs with the unfinished observations that can make a piece distinctive. As Flanagan notes, a promising idea may come from a colleague’s remark or a thought that lingers during a quiet moment. Such beginnings can be rough, but he sees working through them as part of how writers build creative judgment.
“It’s why even ideation with an LLM drags you to the middle.”
Where he draws the line on AI
Flanagan’s post is not a blanket rejection of AI tools. He allows that they can assist with research and offer feedback on structure, particularly for longer pieces. He gives the example of asking a model to review how a draft’s sections and stories fit together.
His boundary is the source material of the writing itself. According to Flanagan, marketers should retain ownership of both the ideas and the language, using tools for support around the work rather than to supply its creative center.
Distinctiveness over polish
That distinction reflects a broader tension for marketers: AI can make content production more systematic, while a highly optimized process may also smooth away the personal perspective that makes a message recognizable. Flanagan’s argument is that imperfect work can be more valuable when it reflects the writer’s own thinking.
“LLM ideas are perfectly created; your ideas are authentically you.”
Flanagan frames writing as a practice that requires repeated effort. Filling a blank page independently, in his account, helps preserve the ability to generate ideas instead of treating every new assignment as a prompt-engineering task.
He closes with a practical recommendation for marketers: use AI selectively for research or structural feedback, but keep authorship of the ideas and words with the human creator. The post presents that restraint as a way to protect creative development as AI writing systems become more capable.
📝 About This Content
This article is based on insights shared by Kieran Flanagan on LinkedIn.
📅 Originally posted on November 3, 2026 | View original post on LinkedIn →