Abstract:

Students in STEM disciplines often face intersecting barriers that limit their success—insufficient time to seek help, lack of effective study strategies, and limited confidence in independent learning. These challenges are particularly pronounced in mathematics, where abstract reasoning and cumulative knowledge demands magnify inequities experienced by first-generation, adult, or working students. Traditional tutoring models assume physical availability and fixed schedules, leaving many learners without timely guidance. This paper introduces an inclusive approach that combines constructivist pedagogy with an AI-enabled tutoring system, CalCompanion, to address both access and learning-strategy barriers in introductory mathematics courses such as Calculus I.

Grounded in the author’s constructivist frameworks—CLEVER (Calculus Learning through Exploration, Verification, Explanation, and Reflection) and TEACH (Transitional, Engaged, Accumulated, Conceptual, Harmonized)—the design promotes critical thinking, self-regulation, and equitable participation. CalCompanion extends instructor presence beyond classroom hours through structured dialogue that mirrors human tutoring. Instead of delivering solutions, the AI tutor poses probing questions, prompts verification, and encourages reflection: “How could you confirm this using a graph?” or “What pattern do you notice?” In doing so, it guides students through cycles of exploration and metacognitive awareness central to the CLEVER model.

This approach directly responds to the digital divide and time poverty experienced by diverse learners. Many students balance employment and family responsibilities that restrict access to traditional office hours or on-campus tutoring. The AI tutor functions as an assistive learning technology—available anytime, fostering persistence, and supporting self-paced study. Embedded reflection tasks and analytics enable learners to monitor progress, identify misconceptions, and receive immediate formative feedback. Early classroom observations indicate that students use CalCompanion to clarify conceptual steps, test reasoning, and reinforce problem-solving confidence when human tutors are unavailable.

By positioning AI as an equitable partner rather than a replacement for human instruction, this project advances inclusive mentoring and tutoring practices in higher education. It models how digital tools can humanize learning while bridging gaps created by time, access, and preparation. The paper discusses design principles, ethical safeguards, and implications for scaling AI-supported constructivist instruction across STEM disciplines. Integrating critical thinking and self-regulation within AI tutoring not only democratizes academic support but also prepares students to become reflective, adaptive problem-solvers in an evolving technological world.

Keywords:

STEM education, mathematics learning, constructivist pedagogy, AI-enabled tutoring, self-regulated learning, critical thinking, inclusive education, digital divide.