2026 ASEE Annual Conference & Exposition

An entry level computing course was redesigned to strategically integrate generative AI
throughout the course to promote inquiry-based learning, conceptual understanding, and reflective
practice. The course balances innovation with caution, teaching students both how to leverage AI
effectively and why developing independent computational fluency remains essential. This paper
describes the pedagogical structure, AI-enhanced assessment strategy, and supporting
infrastructure that together foster deep engagement with computing fundamentals.

The course introduces first-year aerospace engineering students to MATLAB programming
within a design-centered and inquiry-based framework. An overarching goal was to empower
students to explore AI tools as learning companions rather than shortcuts, encouraging curiosity
and critical reflection on their use. The course design emphasizes inquiry-based learning through
active exploration, mastery-based grading to encourage iteration and reflection, and
AI-augmented assessments to enable scalable, personalized feedback in a large-enrollment
setting. The course also integrates asynchronous MATLAB Academy exercises through the
MathWorks–Canvas connection, allowing students to learn at their own pace while freeing
in-class sessions for discussion, debugging, and reflection. Students earn mastery points for
demonstrating proficiency across topics, with opportunities to resubmit work until a threshold of
conceptual understanding is achieved.

The first lab introduces students to generative AI in a low-stakes, curiosity-driven format.
Students write a natural-language “recipe” describing how to compute their final letter grade
using the syllabus. They then prompt the MathWorks AI chatbot to generate MATLAB code that
implements their step-by-step instructions. With only minimal instructor guidance on effective AI
prompting, students engage in a discovery-style activity that highlights both the power and
opacity of AI-generated code. Although most students do not initially understand the generated
code, they revisit it in Week 12 to explain how it works through the lens of their original recipe
after learning more MATLAB syntax and logic. This exercise helps students link
natural-language reasoning with structured programming and reframe code as readable and
interpretable rather than a foreign syntax to be memorized. The activity also sparked significant
excitement early in the semester. Many students independently drew parallels between their
natural-language descriptions and specific code segments, reinforcing the idea that programming
is a communicative process.

To ensure authentic evaluation, major assessments are designed to be resistant to AI-based
solutions. Students must demonstrate conceptual mastery through in-person code interviews and
controlled in-class exams. These assessments require students to explain code behavior, logic,
and structure, emphasizing skills that AI cannot replicate on their behalf. To help students
prepare, an AI-based code interview practice chatbot was developed and hosted on the Texas
Advanced Computing Center (TACC). This optional tool allows students to rehearse verbal
explanations, receive constructive feedback, and gain confidence before their live interviews.
Preliminary feedback indicates that this practice environment improves both understanding and
comfort with articulating technical reasoning. Complementing these authentic assessments, an AI-augmented grading pipeline supports weekly formative evaluation. Short-answer concept quizzes and lab submissions are automatically
exported from Canvas, de-identified for FERPA compliance, and processed through an AI grader
that provides individualized feedback on correctness and reasoning. Human graders then review
and refine this feedback before returning it to students. This hybrid approach maintains human
oversight while enabling scale and personalization. A typical grading cycle for approximately 120
submissions requires roughly three grader-hours and costs about $0.05 in API credits. Both
graders and students have reported that feedback is more specific and actionable than in prior
rubric-based systems.

Later in the semester, students engage in an AI literacy module that contextualizes generative AI
within engineering practice, covering its benefits, limitations, and ethical implications. By this
point, students possess enough technical grounding to critically evaluate how AI assists or
misleads computational reasoning. The final project culminates in a reflective portfolio where
students document their work, highlight selected projects, and share artifacts publicly on social
media or professional platforms. This step reinforces communication skills and helps students
recognize the professional value of computational literacy in an AI-augmented engineering
landscape.

The course redesign has yielded early indicators of improved engagement, confidence, and
metacognitive awareness of AI’s role in engineering education. Students report feeling more
capable of reading and explaining code rather than simply generating it. Future work will
formally evaluate learning outcomes, the effectiveness of mastery-based resubmissions, and the
longitudinal impact of AI-based feedback tools.