Abram Anders stood at the front of a classroom at Iowa State University, watching students hunched over laptops, not just typing—but thinking, pausing, revising prompts, arguing with AI-generated sentences. It wasn’t the effortless writing future some imagined when generative AI arrived. “Writing with AI doesn’t take the work out of writing,” said Anders, associate professor of English and the Jonathan Wickert Professor of Innovation. “It changes it.” That shift is at the heart of a new study published in Computers and Composition, co-authored by Anders and Emily Dux Speltz, an Iowa State alum and assistant professor at Embry-Riddle Aeronautical University. Their findings challenge the widespread assumption that AI makes writing easier by doing the thinking for students—instead, it demands deeper, more deliberate cognitive engagement.
The researchers followed 38 undergraduate students across 22 majors in an experimental “AI and Writing” course over two semesters. What they discovered was a transformation in how students approached writing. Initially, many treated AI like a search engine: type a vague prompt, accept the fluent output. But that approach quickly failed. Students began to realize that good AI-assisted writing required precision, rhetorical awareness, and constant critical evaluation—skills long associated with expert human writers. Three pivotal “threshold concepts” emerged from the study, reshaping students’ understanding of writing with AI.
First, writing with AI is inherently experimental. It requires trying, testing, revising, and trying again. Second, human expertise remains essential. AI may sound confident, but its fluency can mask inaccuracies or shallow reasoning—a trap the researchers call the “fluency trap.” Students had to learn to interrogate AI output, not just edit it, checking claims, refining logic, and aligning arguments with disciplinary expectations. Third, and perhaps most importantly, writing with AI must augment human agency, not replace it. “Generative AI can’t decide what it’s arguing, what matters or why the writing exists,” Anders said. “It’s a tool that requires human direction, judgment and boundaries.”
By the end of the course, students weren’t outsourcing their thinking—they were orchestrating it. They used AI to explore ideas, test arguments, and refine their own reasoning, not to avoid the hard work of writing. This distinction is crucial as AI becomes embedded in academic and professional writing. The study suggests that preparing students for this future isn’t just about teaching them how to use AI tools—it’s about deepening their understanding of writing itself. As Anders put it, “Students must recognize that while AI can generate text, it can’t generate purpose—only the writer can do that.” In a world rushing to automate, this research reminds us that the most powerful tool in writing is still the human mind.