Abstract
Access to effective and equitable education remains one of the most pressing challenges in the United States, particularly in K–12 systems shaped by uneven access to technology, funding, and institutional support. The rapid integration of artificial intelligence into classrooms has introduced both opportunity and risk within this broader problem. On one hand, AI tools promise personalized learning, faster feedback, and expanded educational resources. On the other hand, these benefits are not distributed equally. In Virginia K–12 schools, disparities in device access, broadband availability, and teacher training create uneven conditions for AI adoption. Students in under-resourced districts are less likely to benefit from AI-enhanced instruction, reinforcing existing achievement gaps rather than closing them. This thesis portfolio addresses the broader problem of how emerging technologies can improve educational outcomes without deepening structural inequality. While other factors such as curriculum policy, teacher shortages, and funding disparities also shape this issue, they fall outside the scope of this work. Instead, this portfolio focuses specifically on the technical development of AI systems and the social conditions that influence their impact in real educational settings.
The technical component of this thesis explores the design and implementation of a direct freight marketplace built to reduce costs for independent truckers and small shipping companies by cutting out traditional freight brokers. The project was motivated by the high fees that brokers charge in the hotshot trucking industry, often between 15 and 20 percent of each load's value, which hit owner-operators and small fleets the hardest. The system was developed using Next.js with a Supabase PostgreSQL backend and deployed through a web-based interface where shippers, dispatchers, and drivers can post loads, claim and assign freight, message each other in real time, and track deliveries from pickup to drop-off. Key features include role-based access for both independent drivers and company-affiliated users, a profitability calculator with dynamic state-level fuel pricing, interactive route mapping through OpenStreetMap and OSRM, CDL and company verification systems, and an AI-powered support chatbot. The platform charges 5 to 8 percent per transaction instead of the standard broker cut, passing those savings on to the drivers doing the actual work. Testing focused on the full load lifecycle from posting through delivery confirmation across all three user roles, along with system usability under realistic multi-user conditions. Results showed that the platform could coordinate freight movement between shippers, dispatchers, and independent drivers while giving users transparent cost breakdowns and real-time route visualization. That said, the system depends on getting enough users on both sides of the marketplace to be useful, and the verification workflows for CDL licenses and company credentials are still manual processes that would need third-party API integration to work at scale. The project concludes that technology platforms like this can meaningfully lower friction and cost in freight logistics, but their long-term success depends on building a critical mass of users, earning trust through verification and transparency, and staying grounded in the operational realities shared by stakeholders, including direct feedback from active truck drivers in the hotshot industry.
The STS research paper examines the social and institutional factors that shape the adoption and impact of AI in K–12 education, with a focus on Virginia. The central research question asks how unequal access to technology influences the effectiveness of AI-driven educational tools. Drawing on policy analysis, educational data, and prior research, the paper argues that AI does not operate as a neutral tool but instead reflects and amplifies existing inequalities. Schools with greater financial resources are better positioned to implement AI systems, train educators, and maintain the infrastructure required for consistent use. In contrast, underfunded schools face barriers that limit both access and effective integration. The analysis also considers how decision-making at the institutional level, including procurement policies and curriculum design, shapes how AI is used in classrooms. Using a sociotechnical framework, the paper demonstrates that technological outcomes are inseparable from the social systems in which they are embedded. The conclusion emphasizes that without deliberate policy intervention and equitable design practices, AI adoption in education is likely to widen achievement gaps rather than reduce them.
Together, these projects contribute to a more complete understanding of AI in education by addressing both its technical potential and its social consequences. The technical work demonstrates what AI systems can do, while the STS analysis explains who benefits from those capabilities and why. This combined approach highlights the importance of designing not only effective technologies but also equitable systems for their deployment. Future work should focus on improving the reliability of adaptive learning models while also addressing structural barriers to access, including funding disparities and infrastructure gaps. By aligning technical innovation with social awareness, researchers and educators can move closer to realizing the full potential of AI as a tool for equitable education.
