Enhancing UVA’s Computer Science Curriculum: Bridging the Gap with Full-Stack Development; AI Aesthetics: Ethical Implications For Style, Self-Expression, And Diversity

This Science, Technology, and Society (STS) research paper explores how artificial intelligence (AI) is shifting beauty standards on social media through algorithm-driven recommendations, filters, and digital influencers. As platforms like TikTok, Instagram, and YouTube increasingly rely on machine learning to curate content, they reflect popular trends while also influencing how users perceive attractiveness and self-worth. These technologies play a growing role in defining what is seen, celebrated, and ultimately considered “beautiful” online, blurring the line between digital performance and real-life self-image as AI tools move from technical domains into deeply personal and cultural spaces.
Through discourse analysis and a meta-review of scholarly and online sources, the STS paper explores how platforms like TikTok and Instagram use AI to filter and promote beauty content, often rewarding content that aligns with prevailing trends. These platforms rely on machine learning techniques such as supervised learning, clustering, and reinforcement learning to tailor user feeds. As a result, they can create feedback loops that limit exposure to diverse aesthetics. This algorithmic filtering tends to standardize beauty and subtly pressures users, especially young women, to conform to idealized and often unrealistic norms. Repeated exposure to curated beauty content narrows cultural standards and conditions users to internalize and replicate algorithm-approved appearances, reinforcing digital conformity over individual expression.
The research covers three key sociotechnical issues. First, the homogenization of aesthetic trends, where AI recommendations favor repetitive styles and appearances, limiting creative expression. Second, the reinforcement of Eurocentric beauty standards, as algorithms trained on racially imbalanced datasets often favor light skin and Western features. Third, the psychological impact on users, as beauty analysis tools like ChatGPT and AI-generated models offer appearance-based feedback or set unrealistic visual standards. These systems can affect users' self-esteem and identity, particularly when there is little transparency about their training data or the biases they contain.
Theoretical frameworks from STS, like the Actor-Network Theory and the Social Construction of Technology, help interpret these dynamics. They reveal AI systems as active participants in shaping cultural norms, not merely passive tools. These frameworks emphasize how technologies are embedded within networks of human decisions, values, and social forces, highlighting how machine outputs often mirror the biases and limitations of their creators.
The STS paper calls for inclusive design, transparent AI development, and possible harm mitigation regulations. It recommends tools that visibly flag AI-generated content and advocates for training datasets that reflect diverse racial, gender, and aesthetic identities. By doing so, AI can move beyond reinforcing existing biases and instead support a broader, more inclusive vision of beauty.
This technical paper proposes a structural enhancement to the University of Virginia’s computer science curriculum by introducing a dedicated full-stack development course as a prerequisite to Advanced Software Development (ASD). While UVA’s program excels at teaching theoretical concepts and low-level programming, it lacks sufficient instruction in modern, scalable application development, often leaving students underprepared for ASD and real-world software engineering roles.
The proposal advocates for a course modeled after CS 4260: Internet Scale Applications, drawing from personal experience and supported by industry research. This course would cover essential full-stack topics such as front-end design (JavaScript, DOM manipulation), backend architecture (Node.js, APIs), asynchronous programming, database management (SQL and NoSQL), authentication, microservices, containerization (Docker, Kubernetes), and system resilience. The curriculum is designed to offer students hands-on experience building scalable, production-level applications before taking ASD.
Research shows that many graduates possess strong theoretical knowledge but struggle in industry due to a lack of practical experience. By placing this full-stack course in the fourth semester, after Software Development Essentials and before Advanced Software Development, students would gain the applied skills needed to build integrated, user-facing systems. This restructuring ensures that students enter ASD with a stronger foundation, enabling them to focus on advanced concepts rather than struggling with unfamiliar tools and architectures.
The anticipated impact includes improved academic performance in ASD, stronger technical portfolios, and greater confidence during technical interviews. Students would be better equipped to discuss system design, API integration, and scalability, which are frequently emphasized in hiring processes. Additionally, this course also helps students develop more advanced and comprehensive applications, enhancing their project work and improving their readiness for internships.
The paper concludes by calling for a pilot version of the course to assess its effectiveness, with future work including student feedback, post-internship surveys, and curriculum updates aligned with emerging technologies like cloud computing and generative AI. Over time, this initiative seeks to modernize UVA’s computer science program by aligning academic training more closely with industry demands.

School of Engineering and Applied Science

Bachelor of Science in Computer Science

Technical Advisor: Rosanne Vrugtman

STS Advisor: Karina Rider