Agency, AI, and Intellectual Work
January 8, 2026
In a post responding to Rob Lively’s presentation “Resisting AI for Writing Assignments,” Daniel Plate argues that banning AI in the name of student agency produces a paradox. Lively claims that AI threatens agency by removing struggle, reflection, and rhetorical choice from writing. Plate counters that eliminating AI entirely removes students’ ability to choose how they work, thereby undermining agency rather than preserving it. I agree with Plate’s critique and extend it by arguing that AI can function as a form of cognitive support that enables exploration—particularly in how students identify interests and possible career paths— without replacing intellectual work.
Fair Representation of Lively’s Argument
Lively’s concern is grounded in established composition theory. He argues that writing is not merely a way to express thought but a process through which thought is formed. Activities such as prewriting, revising, and adapting to disciplinary conventions require students to engage metacognitively with ideas. When AI performs these tasks, Lively suggests, students “learn neither of these things,” losing opportunities to develop professional identities through writing.
“Developing some thoughtful reflective writers who make choices is the purpose of creating writing choices in the first place.”
This is a genuine strength of Lively’s argument. Writing does involve productive struggle, and a classroom that encourages students to bypass thinking entirely would fail its pedagogical mission. Lively is right to worry about forms of AI use that substitute output for understanding.
The Agency Paradox Plate Identifies
Plate’s critique does not deny the value of struggle. Instead, it focuses on the internal contradiction of Lively’s proposed solution. Lively celebrates students who voluntarily resist AI use, presenting their resistance as evidence of maturity and agency. Yet he ultimately recommends banning AI for “any writing task” at the institutional level.
As Plate points out, the significance of student resistance lies in the fact that it is chosen. A student who decides not to use AI after weighing its limitations has exercised judgment. A student who avoids AI because it is prohibited has merely complied with a rule. By removing the option to choose, bans eliminate the very condition that makes agency possible.
I agree with Plate that agency cannot be developed through restriction alone. If students are trusted to make rhetorical decisions about audience, tone, and evidence, they must also be trusted to make decisions about tools.
External Evidence: AI as Cognitive Support
Research in cognitive science supports Plate’s position that tools do not inherently eliminate thinking. Studies on intelligent tutoring systems and adaptive learning tools suggest that well-designed AI can scaffold thinking rather than replace it. Danielle McNamara and colleagues argue that effective AI systems support learners by prompting reflection, self-explanation, and strategic decision-making rather than providing finished answers.
McNamara et al. note that cognitive support technologies are most effective when they “encourage active processing and metacognitive monitoring rather than passive reception” (McNamara et al.).
This distinction matters for the AI debate in writing classrooms. AI does not have to function as a shortcut. When used as a prompt, counterargument generator, or exploratory aid, it can require students to evaluate, reject, and revise suggestions. That evaluative work is itself intellectual labor.
Extending the Argument: Exploration and Interests
Plate’s argument becomes even more compelling when considered beyond writing assignments. Students increasingly use AI to explore academic interests, research industries, and understand possible career paths. These activities involve judgment, not automation.
In my own experience, using AI to explore job options has helped clarify interests rather than replace reflection. AI can generate possibilities, but deciding which paths align with my values, skills, and goals requires interpretation and choice. The tool initiates inquiry; it does not complete it.
This process mirrors the rhetorical decision-making emphasized in composition pedagogy. Students must assess relevance, credibility, and usefulness. Far from undermining agency, this kind of AI use exercises it.
Agency as Practice, Not Protection
Plate’s response reframes agency as something students must practice rather than something institutions can preserve by prohibition. Banning AI assumes misuse is inevitable. Teaching students how to evaluate AI assumes misuse is a risk worth confronting.
Traditional academic integrity policies evaluate outcomes rather than regulating process. A blanket ban on AI shifts writing instruction toward surveillance of method rather than assessment of understanding. This risks producing compliance rather than judgment.
Conclusion
I agree with Plate that banning AI in the name of agency ultimately undermines it. Lively’s concerns about struggle and learning are valid, but prohibition is not the solution. Agency requires the possibility of choosing otherwise.
When students learn to decide how, when, and whether to use AI, they develop intellectual judgment that extends beyond writing classrooms to professional and exploratory contexts. Trusting students with those choices is not a threat to education; it is its foundation.