Adaptive AI Tutoring: Bringing CS Education to Low-Resource Classrooms Through AI; Who is the Future Coder? The Self-Sustaining Inaccessibility of K-12 CS Education in the USA 6 views

Imagine a planet where all the driving technology was designed and created by only a small, unrepresentative group of those who lived there: Earth is that planet. Currently, the technological workforce is rather homogeneous, hindering the inclusivity of the technology it creates. One of the major factors directly sustaining this homogeneity is the lack of early access to CS education. Everyone has the right to education, and in this technology-driven world, this includes CS education. Without early access, many communities are not only deprived of opportunities in this field but also of having a voice in creating the technology that shapes tomorrow. The digital divide, the gap in technology and internet access, deeply affects many countries worldwide. Children living in affected areas are unable to access the same level of CS exposure, which hinders their ability to pursue this as a career. As a solution to this, my capstone proposes a lightweight AI tutor that can run on little compute power, run its language model locally, and avoid frequent internet access. The creation of this application would allow children/learners worldwide to have a platform where they can learn basic CS skills interactively and personally, with feedback from their AI tutor. This technology aims to mitigate the inaccessibility of CS education to increase the chances of more diverse technical teams whose innovation can perpetuate equity through the voices of those who create it. Inaccessibility to CS education is caused by a web of economic, social, and political factors that any mitigation strategy must consider. My STS Research paper examines how the inaccessibility of CS education is a self-sustaining feedback loop of inequality in the USA. I collected various scholarly works that identify the actors responsible for causing inaccessibility to create an Actor-Network web. After this, I applied the Social Construction of Technology to understand how different social groups interpret CS education and how that affects its accessibility. For example, affluent families may see CS education as more mandatory because they can afford it, while less affluent families may not feel the same. Finally, I applied Sociotechnical Systems Theory to understand how a feedback loop of inequality is created. The lack of access to CS education leads to lower participation in CS among certain demographics, diminishing inclusivity in the technology created. This leads to less motivation for early access to CS education. In the end, the inaccessibility of CS education widens the skills gap, creating two classes of people: those given the opportunity to understand and pursue CS and those not given this opportunity in the first place. Lessening this gap is imperative for creating the technology of tomorrow, and an inclusive and equitable future depends on a diverse set of creators, something we have yet to achieve.

BS (Bachelor of Science)

Adaptive AI; AI; CS Education; Artificial Intelligence; Digital Divide

School of Engineering and Applied Science Bachelor of Science in Computer Science Technical Advisor: Briana Morrison STS Advisor: Sean Murray

English

All rights reserved by the author (no additional license for public reuse)

2026-05-07