By: Jeff Smith
Artificial intelligence is becoming increasingly integrated into professional practice, and structural engineering is no exception. AI tools can help engineers analyze large amounts of data, monitor infrastructure, optimize designs, identify patterns, and improve efficiency.
But as the technology becomes more capable, it also raises an important question: How can structural engineers use AI without compromising their professional and ethical responsibilities?
The answer begins with a principle that predates AI: technology may assist professional judgment, but it does not replace it.
AI Does Not Change Professional Responsibility
One of the most enduring lessons in structural engineering comes from the Hyatt Regency walkway collapse. The tragedy demonstrated the potentially devastating consequences of design changes that are not adequately reviewed and reinforced the responsibility engineers bear for the work they approve.
AI introduces a new tool into that process, but it does not fundamentally alter the engineer’s responsibility.
An engineer may use AI to analyze data, recommend materials, optimize a design, or identify potential problems. But if the engineer ultimately approves the work, the engineer must be able to understand and support the underlying conclusions.
“The AI recommended it” is not a substitute for professional judgment.
That principle becomes especially important as AI systems become increasingly sophisticated and their recommendations appear more authoritative.
Three Ethical Risks Engineers Should Consider
1. Accuracy Depends on the Data
AI systems are only as reliable as the information they receive and the data on which they were trained.
Consider an AI system used to evaluate the seismic risk of existing structures. If its training data primarily reflects buildings in urban areas, the system may not accurately assess structures in rural communities or infrastructure built under different conditions.
Likewise, an AI-powered bridge monitoring system may identify patterns in sensor and maintenance data that humans could easily miss. But incomplete or inaccurate data could cause the system to overlook a significant defect—or incorrectly assess the seriousness of a problem.
Engineers should therefore understand the limitations of the data behind an AI recommendation and independently evaluate whether the output makes sense for the particular project.
2. Engineers Must Be Able to Evaluate the Recommendation
Some AI systems operate as a “black box,” producing an answer without clearly explaining how that answer was reached.
That presents an obvious challenge in structural engineering.
Suppose AI recommends reducing the size of a structural component because its analysis indicates the change will still satisfy expected loads. If the engineer cannot determine what assumptions were made, what variables were considered, or why the AI reached that conclusion, reliance on the recommendation may create significant risk.
AI can be an effective analytical tool, but engineers should be cautious about relying on recommendations they cannot meaningfully evaluate.
Professional judgment requires more than accepting a sophisticated-looking result.
3. Confidentiality and Cybersecurity Matter
The ethical use of AI also extends beyond calculations and design.
Engineering firms may use AI tools that process project drawings, specifications, infrastructure data, client information, or sensor readings from critical facilities. Uploading that information into an AI platform can create confidentiality, privacy, and cybersecurity concerns.
For example, an AI-powered structural monitoring system could collect detailed information about vulnerabilities in a bridge or other critical infrastructure. If that data were compromised, the consequences could extend well beyond the engineering project itself.
Before using an AI platform, firms should understand how information is stored, whether it is used to train the system, who can access it, and what security protections are in place.
Client information should not be entered into AI systems without appropriate safeguards and consideration of contractual confidentiality requirements.
Practical Guidelines for Responsible AI Use
Engineering firms do not need to avoid AI altogether. In fact, used appropriately, AI can improve efficiency and help engineers identify risks and opportunities that might otherwise be difficult to detect.
The key is establishing reasonable safeguards.
Firms incorporating AI into their practices should consider several basic principles:
- Use AI as a tool, not a decision-maker. Significant engineering decisions should not be based solely on an AI-generated recommendation.
- Independently verify important outputs. Calculations, assumptions, and recommendations should receive appropriate professional review.
- Understand the limitations. Engineers should consider the quality of the underlying data and whether an AI system is appropriate for the particular project.
- Protect confidential information. Firms should establish rules regarding what information may be entered into AI platforms and which tools are approved for professional use.
- Document significant AI use. When AI meaningfully contributes to an analysis or design decision, documenting its role and the steps taken to verify the output may help demonstrate responsible professional judgment.
AI Should Assist the Engineer—Not Become the Engineer
As AI becomes more common, engineers may feel pressure to adopt new tools simply because competitors, clients, or project teams are using them.
But professional ethics are not determined by technological trends.
The same principles that have always governed sound engineering remain relevant: understand the work, question assumptions, recognize limitations, protect sensitive information, and exercise independent professional judgment.
AI has tremendous potential to make structural engineers more efficient and better informed. The ethical challenge is ensuring that its increasing capabilities do not lead professionals to surrender the judgment and accountability that remain uniquely their own.
This article is for informational purposes only and is not intended as legal advice.