AI-Assisted Structural Testing Education: Test Design and Evidence-Based Interpretation

Abstract

This paper proposes an artificial intelligence (AI)-assisted redesign of the Civil Engineering Structural Testing Technology course, with the development of disciplinary judgment as its central objective. The redesign addresses three problems: difficulty applying knowledge learned in separate chapters to specific test decisions, insufficient verification of parameter choices and loading arrangements in student-designed test plans, and a lack of observational evidence to support the interpretation of measured data. In theoretical instruction, knowledge and reference materials are organized around relationships among test decisions; AI assists with organization and retrieval, while instructors and students jointly check the applicability of relevant standard provisions. The open-topic design assignment follows an iterative sequence of initial drafting, AI questioning, student verification, and instructor review, introducing verification before implementation. Laboratory preparation, discussion, and reporting are organized around observational evidence so that interpretations can be traced to original records. Two cases are presented: the design of a shear test for a simply supported concrete beam and an investigation of the static and dynamic characteristics of a uniform-strength cantilever steel beam. The cases examine material parameter selection and the conditions for uniform strength, respectively, and specify AI-assisted tasks, student verification methods, and assessment criteria. The course uses general-purpose large language models accessible to students and requires no dedicated platform or additional hardware investment. Its educational effectiveness remains to be examined using evidence from implementation.

Introduction

The text presents an AI-assisted redesign of the Civil Engineering Structural Testing Technology course. The main goal is to improve students’ ability to make evidence-based engineering judgments, rather than simply following laboratory procedures or producing final reports.

Main problem

The existing course combines theory, open-topic test design, and physical laboratory experiments, but these activities are poorly connected:

• Students design their own test plans but usually do not implement them.

• Laboratory tests use pre-prepared specimens and fixed procedures.

• Students therefore do not verify whether their design choices would work in practice.

• Structural tests are often destructive, so design mistakes cannot easily be corrected after testing begins.

• Instructor feedback alone cannot provide detailed verification of every group's calculations, loading plans, and measurement arrangements.

Proposed solution: AI-assisted course redesign

The paper proposes using AI as a supporting tool, not as a replacement for engineering calculations, experiments, or instructors. AI is used to:

• Organize technical knowledge and reference materials.

• Question students about their test-design decisions.

• Identify missing conditions or inconsistencies.

• Help students examine alternative solutions.

• Support data analysis and interpretation.

• Encourage students to verify AI-generated suggestions using textbooks, standards, calculations, and experimental evidence.

Four major competencies

The redesigned course focuses on:

1. Test-plan design – selecting specimens, loading methods, measurements, and equipment with appropriate justification.

2. Laboratory operation – correctly using instruments and observing structural behavior.

3. Data analysis – processing measurements and connecting conclusions to experimental evidence.

4. Teamwork – coordinating design, calculations, laboratory operations, and analysis. plan, then use AI to identify issues that require verification. They independently check the suggestions through calculations, literature, textbooks, and standards, revise their plans, and explain why they accepted or rejected particular recommendations

Three key teaching approaches

1. Knowledge organized around test decisions
Instead of teaching concepts as isolated topics, the course connects theory directly to testing decisions. Students learn why particular loading methods, instruments, measurement locations, and parameters are appropriate.

2. Iterative test design and verification
Students first prepare a test plan, then use AI to identify issues that require verification. They independently check the suggestions through calculations, literature, textbooks, and standards, revise their plans, and explain why they accepted or rejected particular recommendations.

3. Evidence-based interpretation of experiments
Students are trained to connect observations and measured data to their conclusions. For example, cracking, yielding, strain changes, or unusual spectral peaks must be explained using actual experimental records rather than simply being labelled as "errors."

Laboratory applications

AI-assisted activities are applied to experiments involving:

• Strain-gauge bonding and measurement.

• Flexural testing of concrete beams.

• Shear testing of concrete beams.

• Nondestructive testing of steel trusses.

• Static and dynamic testing of cantilever steel beams.

Students retain original measurements, calculations, processing steps, and observations so that their conclusions can be verified.

Assessment

Assessment focuses on the quality of engineering reasoning, rather than the amount of AI-generated content. Students are evaluated through:

• Test-design assignments.

• Practical laboratory assessments.

• Laboratory reports.

• Examinations.

• Revision records and supporting evidence.

• Team-member contributions.

Students also document their AI use, including the tool used, suggestions received, verification performed, and resulting revisions.

Case study: Concrete beam shear test

One major case involves designing a shear test for a simply supported reinforced-concrete beam. Students determine:

• Specimen dimensions and reinforcement.

• Material parameters.

• Expected load capacity.

• Loading increments and holding periods.

• Measurement locations.

• Equipment and organizational requirements.

A major learning point is distinguishing between measured material properties, design values, and converted material parameters. Students examine how different parameter choices affect predicted failure loads and consequently influence the loading plan.

Conclusion

This paper develops an AI-assisted redesign of the structural testing technology course centered on disciplinary judgment. The redesign organizes theoretical knowledge around test decisions, iterative design around questioning and verification, and experimental interpretation around observational evidence. The grounds for judgment and contributions to teamwork are incorporated into assessment. Teaching content, learning activities, and assessment materials are thus aligned with the same competency requirements. The shear-test design and steel beam experiment cases focus on material parameters and uniform-strength conditions, respectively. They clarify the roles of AI assistance, student verification, and instructor guidance and illustrate how tasks can be designed within the proposed course structure. The course uses general-purpose large language models accessible to students and requires no dedicated platform or additional hardware investment. This paper focuses on the course redesign and case designs. Their effectiveness still needs to be examined through evidence from actual teaching, including revisions to test plans, laboratory reports, and independently completed responses. Future work will use this evidence to refine task arrangements and assessment methods.

References

[1] S. Xiao, “Exploration of AI-Assisted Design Teaching in Undergraduate Earthquake-Resistant Courses”, Advances in Education, vol. 16, pp. 1644–1651, 2026. [2] C. Song and Y. Lü, “Study on Design and Application of the Teaching Model for “Computer-Aided Design” Courses Empowered by DeepSeek”, Printing and Digital Media Technology Research, no. 3, pp. 185–192, 2025. [3] Z. Chen and T. Xu, “Teaching design and practice in a ceramics design course assisted by generative artificial intelligence”, Ceramic Studies, vol. 39, pp. 103–105, 2024. [4] J. Jiang, R. He, and M. He, “Teaching reform for civil engineering structural testing at local application-oriented universities”, Education Modernization, vol. 4, pp. 83–85, 2017. [5] P. Yang, C. Jia, L. Liu, et al., “Research on combining method of engineering experiment and theoretical teaching”, Journal of Architectural Education in Institutions of Higher Learning, vol. 29, pp. 140–147, 2020. [6] B. Qiu, G. Zheng, and W. Luo, “Research on Teaching Reform of Civil Engineering Structural Testing Course Integrating Virtual Simulation Technology”, The Theory and Practice of Innovation and Entrepreneurship, pp. 163–165, 2026. [7] W. Li and S. Chen, “Design and Practice of Civil Engineering Teaching Experiment by Using Innovative Structural Experiment as a Carrier”, Research and Exploration in Laboratory, vol. 45, pp. 185–191, 2026. [8] Y. Hao, Y. Zhou, X. Chen, et al., “Teaching Innovative Study on Structural Testing in Architecture”, Journal of Anqing Normal University (Natural Science Edition), vol. 29, pp. 125–128, 2023. [9] Y. Liu, M. Sun, and Y. Hu, “AI-Enabled BIM Course Reform from the Perspective of Dynamic Competency Maps”, Development Guide to Building Materials, vol. 24, pp. 28–30, 2026. [10] Y. Meng, X. Liu, J. Chen, et al., “Innovative Teaching Model and Practical Research on AI-Enhanced Industry Case-Based Approaches in Chemical Engineering Laboratory Courses”, Chemical Engineering Design Communications, vol. 52, pp. 81–85, 2026.

Copyright

Copyright © 2026 Yaping Ge. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.