2026 ASEE Annual Conference & Exposition

This complete research paper investigates the current level of artificial intelligence (AI) literacy among first-year engineering students by examining how they formulate prompts, interpret AI responses, and apply those responses to engineering problem-solving tasks. In addition to characterizing students' AI use, the study aims to inform instructional practice by identifying specific challenges and opportunities for supporting productive and critical AI use in early engineering courses. While research has documented widespread student adoption of AI tools, less is known about how novice engineering students prompt these tools, evaluate AI-generated outputs, and demonstrate emerging forms of AI literacy. This study examines aspects of AI literacy among first-year engineering students through their interactions with a generative AI tool in a programming-based engineering task. Participants were drawn from an introductory engineering course at a large public R1 university. Students completed a survey describing their AI prompting practices and evaluating a sample AI-generated solution to a familiar engineering problem. In addition, student pairs completed a MATLAB-based programming activity in which use of an AI assistant was permitted and required to be documented through prompt logs and reflective reports. Qualitative analysis of prompt lists, reflections, and survey responses was conducted using an established AI literacy framework, with particular attention to effective prompting and critical evaluation of AI outputs. Results indicate that most first-year engineering students operate at or below novice level of prompting, a component of AI literacy. Common prompting strategies included copying and pasting full problem statements, using AI primarily for debugging, requesting assistance one line of code at a time, or avoiding AI altogether. Students who employed more specific, constrained prompts or used AI to explain errors rather than generate complete solutions reported greater success and less frustration. However, a significant proportion of students were unable to identify errors in AI-generated solutions, highlighting limitations in critical evaluation skills. Together, these findings are synthesized into an actionable set of instructional recommendations for integrating AI tools into first-year engineering courses while maintaining student ownership, disciplinary reasoning, and critical thinking. These findings suggest that exposure to AI tools alone is insufficient for developing productive and ethical AI use in engineering education. Instead, explicit instruction in prompt construction, critical evaluation of AI outputs, and ethical boundaries is necessary, particularly in first-year courses where foundational problem-solving habits are formed. This study provides baseline evidence of first-year engineering students’ AI literacy and offers guidance for integrating AI literacy instruction into early engineering curricula.