Addressing Resource Literacy Through Structured AI Integration in Online English Composition

Synthesis: Erica Reed argues that the deeper instructional problem in online first-year composition is not generative AI but resource literacy — students' ability to identify, evaluate, select, and strategically use academic supports. Drawing on three years of iteration in fully online college courses at Cochise College, she describes the CLEAR cycle (Clarify, Locate, Engage, Assess Feedback, and Revise/Reflect), which treats Generative AI as one optional resource among rubrics, tutors, Librarians, peers, and instructor. CLEAR rests on five design commitments: GenAI is optional, transparent, evaluated, one resource among many, and paired with reflection. Reed reports that students used GenAI mainly for brainstorming and clarification rather than full essay production; that reflection and disclosure reduced overreliance; that comparing AI feedback with rubric criteria strengthened Evaluative Judgment and human Help-Seeking; and that optionality preserved Learner Agency. Her recommendation is design-level: embed task clarification, resource planning, feedback comparison, and structured reflection into ordinary coursework so resource use becomes teachable.

Key Findings

• Resource literacy, not AI use, is the core problem: struggling students were "often not students without access to support, but students who did not consistently recognize when help was needed."
• CLEAR — Clarify, Locate, Engage, Assess Feedback, Revise/Reflect — makes resource use visible: Clarify decodes task demands, Locate maps resources with justification, Engage models prompting, Assess Feedback compares sources, Revise/Reflect closes the loop.
• GenAI is positioned as optional, transparent, evaluated, one resource among many, and paired with reflection — five design commitments aligned with self-regulated learning and universal design for learning.
• Students most often reported using GenAI for idea generation, clarifying assignment expectations, early outlining, and feedback on developing drafts, not for full essay review or final revision.
• Structured reflection plus required AI transparency statements were associated with decreased initial overreliance; earlier iterations documented GenAI use without analyzing whether it helped thinking.
• Modeled strong and weak prompts appeared to shape behavior more than syllabus policy; comparing AI feedback with rubrics also strengthened evaluation of peer, instructor, and tutor comments.

What resource literacy means when a tool answers back

Resource literacy is Reed's term for students' ability to identify, evaluate, select, and strategically use academic supports in ways that advance their learning goals. In online composition that means interpreting prompts accurately, consulting rubrics and examples, seeking feedback, using library and tutoring services, and judging whether feedback is useful for revision. Reed grounds it in Self-Regulated Learning: task value, Self-Efficacy, metacognitive monitoring, effort regulation, and help-seeking. Her recurring observation was that students who struggled were not those without support but those who did not recognize when help was needed, which resource fit the task, or how to judge feedback once received. Generative AI did not create that difficulty; it made it visible, because fluent output is easily mistaken for authoritative support. The instructional question shifts from whether students may use AI to whether the course teaches AI Literacy.

How the CLEAR cycle structures integration

CLEAR — Clarify, Locate, Engage, Assess Feedback, Revise/Reflect — ran in online first-year composition at Cochise College, grounded in five design commitments: GenAI is optional, transparent, evaluated, one resource among many, and paired with reflection. In Clarify, students restate assignments, name the rhetorical goal, and list evaluation criteria in a weekly journal, with GenAI modeled as a questioning tool rather than a drafting tool. Locate has them name the hardest part of the assignment, select supports, and justify the fit; students choosing GenAI also define its intended function. Engage supplies recorded modeling contrasting weak prompts against stronger ones. Assess Feedback requires comparison across AI, peer, instructor, and tutor comments. Revise/Reflect closes the cycle with Metacognition about which supports helped or hindered progress.

What students actually did across three academic years

Reed reports five recurring outcomes and labels them course-based observations rather than generalizable findings. Students most often used GenAI for idea generation, clarifying assignment language, early outlining, and feedback on developing drafts rather than full essay review. Heavier drafting use produced writing polished in tone but weak in specificity, and comparing drafts against rubric criteria exposed generic phrasing and misalignment, making the rubric a corrective mechanism. Structured reflection and required transparency statements were associated with decreased overreliance. Modeled prompting seemed more influential than policy text, and comparing GenAI feedback with rubrics produced what Reed calls productive skepticism, redirecting students toward human feedback even when it took more time. The most significant shift, she writes, was conceptual: CLEAR made resource decisions visible and analyzable rather than eliminating misuse.

What this means for practice

• Teach resource selection explicitly: require students to justify which support fits a task rather than assuming they will seek help appropriately.
• Build CLEAR-style routines — clarification prompts, resource mapping, feedback comparison, structured reflection — into assignment modules as standard design.
• Model prompting and critique in lectures and written examples; policy language alone did not appear to change behavior.
• Keep GenAI optional alongside human supports; students who opted out completed the cycle with tutors, librarians, peers, and instructor feedback.

Limitations

• Findings are course-based observations from one instructor at a single institution, not a controlled study; Reed states they are not generalizable.
• No participant counts, comparison groups, or statistical measures are reported; the only figures are three academic years and five outcomes.
• The model is described at the cycle level; assignments, rubrics, and journal prompts are not reproduced, so replication requires interpretation.

Connected Concepts

• AI Literacy
• Self-Regulated Learning
• Universal Design for Learning
• Evaluative Judgment
• Metacognition
• Prompt Engineering

Connected Articles

• "Young Scholar[s] on the Beat": Multimodal Composition as a Form of Critical AI Literacy Pedagogy — Critical AI literacy built through multimodal composition in a writing classroom.
• Scaffold or Shortcut? Postgraduate IT Students' Use of Generative AI and Self-Regulated Learning — Whether GenAI acts as scaffold or shortcut for self-regulated learning in postgraduate work.
• Efficiency vs. Effectiveness: Self-Regulated Learning with LLM-Mediated Help-Seeking — How LLM-mediated help-seeking trades efficiency against learning effectiveness.
• "Should I Tell My Teacher?" Student AI Disclosure Practices, Stigma, and Self-Regulated Learning in Higher Education — Student AI disclosure practices, stigma, and their link to self-regulation.

Citation

Reed, E. (2026). Addressing Resource Literacy Through Structured AI Integration in Online English Composition. Journal of Instructional Design and Technology, 1(2), 37-44.