Transforming Structural Engineering: Embracing the AI Revolution

To view the figures and tables associated with this article, please refer to the flipbook above.

Artificial Intelligence (AI) has swiftly transitioned from a distant futuristic concept into an integral component of engineering practice. In civil engineering—and particularly structural engineering—AI now offers significant opportunities to improve efficiency, accuracy, and reliability in both routine operations and complex analytical tasks. This article discusses current AI structural engineering applications, future possibilities, and critical considerations for the profession.

AI technologies such as computer vision (which enables computers to interpret visual data), natural language processing (NLP, which allows machines to understand and generate human language), and machine learning (ML, which uses data to improve performance without explicit programming) are increasingly integrated into structural engineering workflows. These technologies enable automation and improved precision across several critical tasks, from infrastructure inspections and code interpretation to quality assurance and control.

One notable area where AI is currently demonstrating practical value is inspections. Traditional inspections, reliant on manual visual checks and tedious documentation, are labor-intensive and prone to human error. AI-powered computer vision now facilitates rapid, automated analysis of visual data from drone imagery, satellite photos, or stationary cameras. For instance, when trained within specific domains, AI can swiftly detect structural anomalies, including cracks, corrosion, and deflections, and classify damage severity. These capabilities help engineers prioritize maintenance activities, drastically reducing time and cost. Similarly, AI-powered predictive models now help structural engineers forecast potential failures and schedule preventive maintenance. This approach relies on analyzing historical data, structural characteristics, environmental conditions, and real-time data from embedded sensors to predict and avert structural issues before they become critical, thereby significantly prolonging the lifespan of infrastructure. For instance, an AI model can continuously process vibration and strain data from sensors on a highway bridge to learn its normal structural behavior. Once this behavior is understood, the model can detect subtle changes in dynamic response patterns to predict when metal fatigue might surpass safety limits. This allows engineers to plan necessary retrofitting, such as replacing FRP or steel plates, or imposing load restrictions, long before cracks appear.

AI-driven structural design optimization is another critical area of interest as structural engineers frequently encounter complex, multi-objective design scenarios that need to balance safety, economy, sustainability, and performance. AI algorithms, such as genetic algorithms and neural networks, can quickly generate and assess thousands of potential design configurations, identifying optimal or near-optimal solutions that human engineers might overlook due to complexity or time constraints. While these tools are currently limited to use in early design phases and are available to only a few firms, the skills and capabilities are spreading.

AI-enhanced structural health monitoring (SHM) offers continuous monitoring of structures using distributed sensor networks—collecting data on strain, vibration, temperature, and other variables. Traditional SHM methods rely on periodic checks and offline data processing. In contrast, AI systems detect real-time deviations and pattern changes that may indicate structural damage. These systems are being built and trained as both engineer-in-the-loop systems reduce the time to notify the asset owner of issues but could be used as a direct notification tool, massively reducing the level of staff effort from the status quo.

Such systems are especially valuable in seismic or high-wind regions. In California and Japan, AI-powered SHM platforms have been used to assess buildings immediately after earthquakes, supporting faster decision-making and emergency response. As these systems evolve, they are transitioning from passive reporting tools to active decision-support platforms capable of triaging risk and dynamically adjusting maintenance plans. This represents a fundamental shift from passive to active monitoring, empowering engineers and infrastructure managers with unprecedented insight into real-time structural behavior.

While inspections are an immediate benefit, developing, interpreting, and applying building codes and standards is another area ripe for AI-driven transformation. Engineers often navigate lengthy, intricate regulatory texts, a process vulnerable to misinterpretation or oversight. AI—particularly NLP—offers opportunities to assist engineers by parsing complex regulatory language and offering targeted answers. Tools based on GPT-like language models could eventually allow engineers to ask, “What is the live load requirement for a storage mezzanine?” and receive precise, context-specific responses.

While today’s models still struggle with ambiguous or context-sensitive language, improvements are coming to the baseline model’s ability to parse more complicated formatting such as the code-ubiquitous table. That is, they are learning how to hold more fingers in the code book! However, beyond waiting for the models to improve their ability to read our existing code books, embedding AI friendly formatting and logic directly into standards, developing flexible licensing models, and offering user-customizable tools may also represent new opportunities for standards organizations and their industry partners. NCSEA has released a tool called SE GPT based on a database of content in NCSEA webinars and STRUCTURE magazine content. While preparing this article, the authors have had a sneak peek of a similar ASCE initiative to provide a custom GPT interface for some of its content. This is just the beginning of tapping into the industry-wide corpus of knowledge.

The potential of AI to streamline Quality Assurance and Quality Control (QA/QC) processes is enormous. Engineering projects involve vast amounts of documentation, calculations, and data, all of which must be verified meticulously to maintain compliance and structural integrity. AI algorithms can automatically analyze large datasets to identify discrepancies, errors, or deviations from established standards, significantly enhancing reliability and consistency while reducing the manual effort required from engineers. The AI tools on the market today already allow an engineer to upload a reinforcement schedule and shop drawing set and ask AI to do a first pass cross-check review including an instance count and summary table. While this is nowhere near a complete review task, it takes a few seconds, helps the engineer assess how long the complete review may take, and draws attention to areas of concern. This type of task can be done without customization, without training, or without prompt engineering. Using AI as a preliminary “review-before-the-review” can help a senior engineer preparing for a QA/QC of a project or a junior engineer preparing to complete a review of external partner’s work. Such capabilities could allow structural engineers to reallocate their attention from routine checks to more critical design and decision-making activities, optimizing overall project efficiency and closing information gaps by cross checking basis of design, drawings, specifications, and calculation packages.

Achieving this type of QA/QC requires connecting several different AI technologies in the piecewise approach (large language models may parse specifications and emails, computer vision may be required to review drawings, and machine learning solutions would be used to analyze outputs from BIM and structural analysis models). Each solution will require a significant degree of firm-specificity that will not be addressed by vendors; thus, efforts within firms will be required to solve issues specific to their own workflows/risk patterns.

Beyond enhancing safety and reliability, AI also presents immense opportunities to address sustainability challenges within structural engineering. As global awareness about environmental impact grows, engineers are increasingly tasked with creating structures that meet stringent performance criteria and minimize ecological footprints. AI algorithms are particularly effective at optimizing material use and reducing waste, directly contributing to sustainability goals.

Here, material optimization using AI involves sophisticated techniques such as topology optimization and generative design, where algorithms iteratively explore countless configurations to find the most efficient structural forms. Unlike conventional design methods, AI-driven generative design rapidly evaluates and compares materials, shapes, and structural layouts, automatically considering environmental impact metrics such as embodied carbon, energy consumption, and material recyclability. For example, AI tools have been developed that connect key design criteria such as geometry, loading, materiality, vibration, and embodied carbon takeoff into a single live interactive interface. This directly translates into lower greenhouse gas emissions, reduced material extraction, and cost savings.

Additionally, AI can assist engineers in assessing the lifecycle impacts of structural materials more comprehensively. By combining historical data with predictive models, AI-powered assessments provide accurate forecasts of maintenance needs, durability, and environmental impacts over the entire lifecycle of structures, promoting truly sustainable and resilient engineering practices. For example, one of Thornton Tomasetti’s AI-powered structural design tools, Asterisk, allows for rapid design iteration by incorporating geometry, wind and seismic loading criteria, vibration criteria, and material customization; it then allows the engineer to immediately capture outputs such as member sizes, structural quantity, and embodied carbon takeoffs.

Despite these advancements, the widespread and effective integration of AI into structural engineering practices demands that several key challenges be addressed. First is the need for clearly defined guidelines and ethical frameworks for AI usage. Organizations such as ASCE, SEI, and NCSEA are instrumental in developing these essential standards. For instance, SEI might consider creating new standards specifically designed to govern AI-driven designs, akin to performance-based design standards and model validation frameworks already established in sectors such as marine safety and fire engineering. Such frameworks would set a clear minimum standard of care, ensuring that AI-enhanced designs achieve the necessary levels of safety, accuracy, and reliability.

While these frameworks are developed, we can lean on item 1.h of the ASCE Code of Ethics that contains a clear reminder to all of us as we seek to incorporate these new tools into our work: “consider the capabilities, limitations, and implications of current and emerging technologies…” this is a reminder that in all of the examples here, the AI tools are proposed as partners, not replacements for the engineer. The time saved through efficiencies gained should be spent focusing on design fundamentals, deeper QA/QA, and ensuring that we are best solving our client’s and society’s needs.

Additionally, the procedures by which standards themselves are developed and disseminated could be significantly enhanced by AI technologies. Standard development typically involves extensive stakeholder collaboration, detailed record-keeping, and administrative processes, all of which consume considerable time and resources. AI can dramatically reduce the time required to synthesize stakeholder feedback, summarize comments, and manage administrative tasks such as generating meeting minutes. Moreover, AI-enabled tools could readily detect and highlight changes between code versions, simplifying engineers' ability to stay updated. By proactively developing customized AI interfaces—such as ASCE/SEI-specific chatbots—engineers could interact directly with standards via intuitive queries, significantly improving accessibility and comprehension.

Embracing AI also demands an educational and professional paradigm shift. Structural engineers' training—both academically and professionally—needs rapid adaptation to equip engineers effectively. Some academic institutions are already integrating "micro" educational programs designed to help students quickly adapt to emerging technological trends. Firms and professional licensing bodies must similarly respond by incorporating AI competencies into certification and training frameworks, recognizing that while it is impossible to master all AI technologies, engineers must become proficient in the tools most relevant to their specific roles.

Furthermore, bridging generational divides within structural engineering organizations is paramount to effectively adopting AI. Younger engineers entering the profession generally possess greater familiarity with digital technologies and AI applications, while senior engineers bring essential depth in practical experience and engineering judgment. Encouraging cross-generational collaboration ensures that organizations use both innovative technological solutions and time-tested engineering ability, fostering a robust integration of AI into practice.

Immediate actions structural engineers should take today include actively engaging with accessible AI tools, such as ChatGPT, to gain familiarity with basic AI capabilities and to develop a sense of what AI can do and what it can’t. Although such tools represent only a fraction of AI's broader potential, routine use can foster increased comfort, creativity, and productivity in daily tasks. Engineers should also prioritize learning and adopting AI-driven technologies directly relevant to their practice, staying current with offerings from organizations like NCSEA and ASCE.

The structural engineering profession stands at a pivotal moment in the integration of AI. While the potential for enhanced efficiency, insight, and innovation is substantial, realizing these benefits requires deliberate, ethical, and collaborative implementation. We must keep our ethical responsibility central in our mind and consistently have an engineer-in-the loop as the new systems are developed. Engineers must work alongside governing bodies and educators to shape standards, define accountability, and cultivate the skills necessary for this new era.
With thoughtful leadership and strategic investment, AI will not replace the engineering profession, it will amplify its impact. Structural engineers who engage early, skill up, and lead responsibly will be at the forefront of building a safer, smarter, and more sustainable future. ■

About the Authors

Kristopher Dane, D.Eng., CPEM, is Associate Principal at Thornton Tomasetti.

M. Z. Naser, Ph.D, PE, is Assistant Professor, at Clemson University and AI Research Institute for Science and Engineering.