Engineering Math with AI: Algebra Sense-Making in First-Year Engineering and Problem-Solving Courses

We report results from an introductory engineering math and problem-solving activity used in two first-year courses: College Algebra and Trigonometry for Engineers & Scientists and Engineering Problem Solving. Students first solved a classic distance–rate–time task by hand (Part A), then used ChatGPT as a tutor to re-solve and explain the problem (Part B), and finally wrote a reflection on what they learned and how AI helped (Part C). A rubric assessed mathematical correctness, clarity of algebraic reasoning with units, and quality of AI interaction and reflection (200 points total).Preliminary findings show that most students reached the correct round-trip times (about 90%). Without explicit prompting, recurring issues included averaging speeds, dropping units, and skipping the initial time = distance ÷ speed setup. When students prompted the AI to explain in steps, check units, and verify answers, their work showed clearer structure and fewer errors than prompts that simply asked for the answer. Across reflections (n≈24 to date), students rated AI helpfulness at about 3.9 out of 5 on average (range 2–5). Learners reported that better prompts improved sense-making, while overly procedural explanations remained confusing unless they requested conceptual justification.We will share the assignment, rubric, and a prompt toolkit instructors can adopt immediately. Implications center on designing AI-supported tasks that cultivate algebraic reasoning, productive prompting habits, and metacognitive reflection, moving AI use in engineering mathematics beyond answer-getting.