This study examined the impact of scaffolded writing instruction supported by artificial intelligence tools on the academic writing development of first-year college students from low socioeconomic and linguistically diverse backgrounds. The research was conducted in multiple sections of freshman English composition courses at a historically Black college. Many students entered the course with limited experience in academic writing, challenges in reading comprehension, and minimal confidence in research-based writing tasks. The instructional design integrated children's literature as foundational texts to reduce writing anxiety and build conceptual understanding of thesis development, paragraph structure, and evidence integration. Artificial intelligence tools were used as guided supports for brainstorming, organizing ideas, clarifying assignment expectations, and providing formative feedback during the drafting process. AI use was intentionally framed as a coaching tool rather than a replacement for student thinking. Data were collected through rubric-based writing assessments, student reflections, classroom observations, and comparisons of early and final writing samples. Findings suggested that scaffolded instruction combined with responsible AI integration supported improved thesis clarity, stronger use of textual evidence, and increased student confidence in academic writing. Students also demonstrated greater engagement with the writing process and a clearer understanding of revision as a learning practice. The study contributed to ongoing conversations about equitable writing instruction, ethical AI use in higher education, and pedagogical strategies for supporting underprepared college writers in achieving academic success.
Implementation steps and strategic initiatives
Amanda Haywood-Cotton's work at a historically Black college calls for an immediate, structured curriculum redesign of freshman English composition courses. The first step is to convene a faculty working group that includes English instructors, instructional designers, and student success advisors to co-develop a revised course framework. This framework must embed culturally relevant children's literature as scaffolding texts, pairing each literary selection with explicit writing instruction targeting thesis development, paragraph structure, and evidence integration. A curated list of titles representing diverse voices and low socioeconomic contexts should be assembled and reviewed for pedagogical alignment before the next academic term.
Simultaneously, the institution should identify and pilot two or three AI writing-support tools — such as Grammarly, Turnitin's AI feedback features, or purpose-built coaching platforms — that can be integrated into the LMS without requiring students to create external accounts. Clear ethical use guidelines must be co-created with students, emphasizing that AI functions as a drafting coach rather than a ghostwriter. Faculty should model this framing explicitly during the first week of class, using think-aloud demonstrations that show how AI suggestions are evaluated and accepted or rejected by the student author.
The scaffolded instructional sequence should progress in three stages: (1) guided practice with children's literature and AI brainstorming tools; (2) structured drafting with AI feedback and peer review; and (3) independent revision with instructor conferencing. Each stage should be supported by detailed instructor guides, sample student work exemplars, and formative checkpoints that allow early identification of students who need additional support. A mid-semester review meeting of the faculty working group should assess progress and make real-time adjustments to pacing and resources.
By the end of the first full implementation cycle, the institution should produce a documented instructional model that can be shared with peer HBCUs through the SMART Global Technology Innovation Center network. This dissemination plan should include an annotated course syllabus, a faculty facilitation guide, and a student-facing AI use policy template. Presenting findings at regional and national conferences on HBCU pedagogy will further amplify the model's reach and invite critical peer feedback that strengthens subsequent iterations.