Prof. Judgment Before Answers: Generative AI in First-Year Mathematical Analysis
DOI:
https://doi.org/10.22158/jecs.v10n4p37Abstract
Generative AI now produces plausible solutions to many standard exercises in first-year Mathematical Analysis, yet the same systems mishandle quantifier scope, omit essential hypotheses, and invite students to substitute fluent text for mathematical judgment. This paper argues that the pedagogical question is not whether GenAI should enter the analysis classroom, but how its use can be designed so that the student, rather than the tool, performs the cognitive labor the course exists to cultivate. Drawing on recent evidence concerning AI in proof-based undergraduate courses, the paper develops a three-principle design framework—tiered access, error-centered inquiry, and judgment-oriented assessment—and instantiates it through worked examples from the first-year syllabus. The examples span limit computations ( and arguments, rationalization, boundedness), integral computations (substitution, integration by parts, symmetry, the Riemann integral and its limitations), and the convergence of function sequences. In each case the AI output is treated not as an answer key but as a case to be diagnosed, and the student's task is to supply the judgment that distinguishes a valid argument from a plausible one.

















