Forecasting Breakthroughs: Identifying Future Leaders in the Semiconductor Industry; Exploring the Contemporary Effects of Generative AI on the Financial and Marketing Industries and the Risks Embedded in its Utilization

Kuzneski, Ethan, School of Engineering and Applied Science, University of Virginia
Earle, Joshua, Engineering and Society, University of Virginia
Porter, Michael, Systems and Information Engineering, University of Virginia

Generative AI, a cutting-edge technology, has swiftly become a transformative force in various sectors, profoundly impacting the job market in the United States. This STS paper delves into the applications of generative AI and its repercussions on employment, particularly in the financial and marketing industries. Generative AI, distinguished by its capacity to create diverse content, is poised to revolutionize workflows across industries. However, its integration raises concerns about job displacement and workforce inequalities, underscoring the urgency of understanding its implications. Employing Actor Network Theory (ANT) and Winner’s Ideology, this study examines the interplay between generative AI and the financial/marketing sectors. ANT illuminates the roles of entry-level employees, corporate executives, and AI systems, while Winner’s framework underscores the political dimensions of technological artifacts. Generative AI exhibits significant capabilities, from ChatGPT’s conversational prowess to models' ability to generate authentic content. Its applications in marketing (e.g., content creation, personalization) and banking (e.g., customer support, data analysis) promise substantial productivity gains and revenue boosts. However, concerns over bias, privacy, and accountability loom large, necessitating regulatory intervention. Adapting to generative AI is crucial for employees, who must harness its potential to remain competitive. While automation threatens certain tasks, the creation of new roles and skill requirements offers hope for workforce resilience. Yet, ethical considerations, including bias mitigation and privacy protection, demand concerted regulatory efforts to ensure responsible AI development and deployment. Future research must monitor employment dynamics and evaluate the efficacy of upskilling initiatives to address evolving job market demands. Moreover, inclusive development strategies are imperative to mitigate disparities arising from AI adoption. Collaborative efforts among researchers, policymakers, and industry leaders are essential to foster ethical AI practices and maximize societal benefits while minimizing risks. That being said, generative AI presents a transformative yet nuanced landscape for the job market. While its potential for innovation is vast, proactive measures are essential to navigate its ethical and socioeconomic implications effectively.

From the technological rise of generative AI, semiconductors have made their own technological disruption due to their capabilities to enhance computational speed and power. This technical paper explores the possibility of anticipating such disruptions and focuses on constructing a binary classification model to forecast the success or failure of semiconductor startups within five years of their initial funding round. The primary objective is to develop a binary classification model predicting the success of semiconductor startups based on achieving a valuation of over $500 million within five years post-funding. This threshold is chosen for its significance in investment opportunities. The model utilizes various metrics, emphasizing those indicative of future success. The study integrates network analysis, feature engineering, and predictive modeling. Stakeholder networks and relevant metrics are identified, encompassing CEO attributes, investor patterns, and regional headquarters locations. The models are trained on data from PitchBook and extra features were created using OpenAlex. Two types of predictive algorithms were used: Lasso logistic regression and XGBoost. Evaluation metrics, including Area Under the Curve (AUC) scores, reveal the performance of the models. While the logistic regression model achieves an AUC score of 77.68%, XGBoost attains 62.86%. The models highlight significant variables such as workforce size, funding, and company location, suggesting their predictive relevance. It was also found that Chinese semiconductor companies exhibit superior success rates, attributed to government initiatives and economic policies driving industry development. The research underscores the importance of early-stage funding and identifies actionable insights for government policies, investment decisions, and startup strategies. Further research avenues include exploring emerging markets and enhancing model agility through real-time data integration and increased sample size. This report offers valuable insights into forecasting semiconductor startup success, highlighting the significance of early-stage funding and key predictive variables. While limitations exist, continuing to collect more data will allow for the models to become more accurate and applicable in the semiconductor industry.

BS (Bachelor of Science)
Semiconductors, Binary Classification, Machine Learning, Generative AI, Job Market

School of Engineering and Applied Science

Bachelor of Science in Systems and Information Engineering

Technical Advisor: Michael Porter

STS Advisor: Joshua Earle

Technical Team Members: Robert Brozey, Carter Dibsie, Adam Rogers, Lauren Sullivan, David Underwood

All rights reserved (no additional license for public reuse)
Issued Date: