Abstract
The capstone research will fill in the systemic gap of an inconsistent standardized safety education for young people. Firearms are in over 40% of American households. Many of these households store weapons unlocked and loaded in their homes. This creates a hazardous environment for children who lack prior knowledge of proper gun safety protocols. For instance, in November 2025, a four-year-old boy sustained life-threatening injuries after accidentally shooting himself with an unsecured firearm he found in his home in North Carolina. In February 2026, a six-year-old boy was found dead from a gunshot wound to the face after officials reported that he had accidentally shot himself on a Tuesday afternoon at a residence in Montgomery County. The lack of a standardized safety education for these environments has led annually to preventable child accidental injuries and fatalities like these. Considering the human and social dimensions of this technology is important because even a well-designed AI training tool can fail if parents, communities, or institutions do not trust or accept it. Cultural attitudes toward firearms, parental willingness to expose children to the training, and institutional support from schools will all determine whether the technology is actually adopted and effective in reducing child firearm fatalities.
The technologies used to address inconsistent standardized safety education for young children will include robotic systems and artificial intelligence. A Nao humanoid robot equipped with AI-driven natural language processing will first verbally train the child on the four gun-safety steps: "Stop, Don't Touch, Run Away, and Tell an Adult." These four steps will then be assessed through an AI-powered AR simulation that provides instructions, praise, and corrective feedback to guide the child in correctly following the sequence. Both the Nao robot and the AR system's AI will be implemented using retrieval-augmented generation (RAG), which generates responses from a database of relevant firearm safety research. By combining robotic systems and artificial intelligence, the project aims to deliver standardized youth firearm recognition, safety training, and assessment.
Teaching with AI has become more relevant within higher education, which is why the STS research portion will focus on the integration of generative AI in Higher Education. Furthermore, this research will be analyzed within the STS theory of the Social Construction of Technology. SCOT states that technology is heavily influenced by social, cultural, and political factors and not just the result of advancements in technology. For instance, the creation of the early bicycle was heavily influenced by interpretive flexibility, as it was seen as dangerous by women but enjoyable by men; relevant social groups, such as sport cyclists versus touring cyclists; and eventual closure, as by the late 1890s, widespread agreement had formed around the safety bicycle. In the same way, AI in higher education demonstrates interpretive flexibility, as it is seen as dangerous by professors but useful by students; relevant social groups, including professors, students, and institutions; but there is a lack of final closure, as many institutions have yet to agree on the proper use of AI in higher education. This STS framework of SCOT can help navigate the complex growth of AI in higher education.
The method of conducting the STS research will be accomplished by analyzing existing scholarly literature and peer-reviewed journal articles. The sources will be drawn from academic databases such as Google Scholar and ASEE.
The STS research found that GenAI is socially constructed in higher education through how professors, students, and institutions perceive and use it, with no consensus or closure yet reached among these groups. Professors primarily use AI for assessment, grading, feedback, and course design to improve efficiency and personalize learning. Students hold mixed attitudes, recognizing AI as a learning aid while raising concerns about overreliance and academic dishonesty. Institutions have largely failed to provide sufficient training, leaving AI use inconsistent and individually determined.
The Engineering Capstone and STS research will interact by shaping and constraining each other. The Robotic AI firearm-safety training tool will determine how children learn safety behaviors, while the STS analysis of how AI is socially constructed in higher education predicts how society might interpret, trust, adopt, or reject a new AI-powered learning program. If society remains skeptical of AI in educational settings, as the STS research suggests is currently the case, then the capstone's robotic training agent may face resistance from parents, schools, and institutions regardless of its technical effectiveness. Conversely, if the capstone succeeds in demonstrating that AI can reliably teach critical safety behaviors to children, it could help build broader social trust in AI-driven educational tools.
