Abstract
Scientific research increasingly depends on advanced laboratory techniques that require precision, repeatability, and specialized expertise. In materials science laboratories, the transfer and alignment of two-dimensional (2D) materials such as graphene remains a slow and highly manual process that relies on tacit human skill under microscopes. At the same time, artificial intelligence and laboratory automation are transforming scientific work by automating repetitive tasks, augmenting human decision-making, and reshaping the organization of research labor. These changes create both opportunities and challenges. Automation can improve experimental efficiency, reduce human error, and expand access to advanced research tools, but it may also alter training pathways, redistribute expertise, and reinforce inequalities between well-funded and resource-constrained institutions. My technical and STS research projects both address this broader challenge: how AI-assisted laboratory automation can improve scientific research while simultaneously transforming the work of researchers themselves. Together, these projects examine both the engineering feasibility of low-cost laboratory automation and the societal implications of integrating artificial intelligence into scientific practice.
My technical project, AutoStack, focused on the development of a low-cost automated transfer and alignment system for 2D materials. The project addressed the difficulty of manually stacking microscopic materials with micron-level precision, particularly in constrained laboratory environments such as glove boxes or vacuum chambers. The final system integrated three major subsystems: an XY positioning stage, a Z-axis pressure-control mechanism, and a computer-vision alignment module. The platform used lead-screw-driven stepper motors capable of approximately 6.25 µm step resolution, enabling reliable micron-scale positioning while remaining within a total budget of approximately $572. The Z-axis incorporated a pressure sensor and feedback mechanism that halted motion when force thresholds were exceeded, protecting delicate samples during transfer. Additionally, the computer-vision subsystem used edge-based image processing to detect and identify material flakes under a microscope and compute positional relationships between them. Testing demonstrated reliable positioning accuracy, smooth remote operation using an Xbox controller, and successful flake detection across microscope images. By reducing repetitive manual alignment tasks and increasing experimental reproducibility, the project demonstrated that advanced laboratory automation can be implemented at substantially lower cost than commercial systems, potentially expanding accessibility for smaller laboratories and educational institutions.
My STS research project examined how AI-assisted laboratory automation reshapes the roles, skills, and organization of scientific work. Using survey evidence from Nature and Pew Research Center alongside qualitative discussions from online research communities, the study analyzed how researchers perceive the growing integration of artificial intelligence into laboratory workflows. The findings suggest that AI is not simply replacing scientific labor, but instead reallocating tasks and transforming the types of expertise that are most valued within research environments. AI tools increasingly automate repetitive analytical and computational tasks while shifting human researchers toward supervisory, interpretive, and integrative roles. However, the study also identified concerns regarding skill erosion, unequal access to advanced technologies, and growing pressure for researchers to adopt AI tools in order to remain competitive. Drawing on the STS framework of mutual shaping, the paper argues that the effects of automation are not technologically predetermined; rather, they emerge through interactions between technical systems, institutional incentives, and human decision-making. As laboratories adopt AI-assisted automation, they simultaneously reshape expectations surrounding productivity, training, and scientific expertise. The research concludes that AI-assisted laboratory automation represents a transformation of scientific work rather than a replacement of researchers, with important implications for future inequality, education, and research organization.
Together, these projects contribute both technical and social insight into the future of AI-assisted scientific research. The technical project demonstrates that low-cost automation systems can meaningfully improve laboratory precision and accessibility, while the STS research highlights the broader institutional and labor implications of integrating such systems into scientific practice. Although the AutoStack system remains an early-stage prototype, the project established a functional foundation for future development in automated materials research. Likewise, the STS analysis suggests that the long-term consequences of laboratory automation will depend not only on advances in artificial intelligence, but also on how researchers, universities, and funding institutions choose to integrate these technologies into scientific work. Collectively, these projects emphasize that engineering design and societal outcomes are deeply interconnected, and that the future of laboratory automation must be understood as both a technical and sociotechnical transformation.
