We use our experiences to inform our future. That’s the essence of an engineer’s superpower of creative problem solving.
Back in 1981, AE offices had several receptionists or typists and a whole room of drafters. First came the Magcard machines, which replaced the electric typewriters. I remember a typist at retirement age quit, as she felt “that machine would steal her soul.”
If you were doing Department of Transportation work, the new standard was an Intergraph workstation—at a cost of $200,000 to $250,000. DOTs were paying contracts to underwrite the investments, as they were clearly not regular “tools of the trade.” There was panic about all the jobs that would be lost due to these new tools.
As time moved on, we moved from CADD to 3-D modeling to digital twins. We didn’t lose jobs, we added them—but needed to re-skill our workforce. As a result, we better optimized solutions, better coordinated plans and generated better construction drawings.
Another New Technology
Fast forward to today, where the mere mention of AI stirs fear in many hearts. Governments, businesses, schools, professional and trade organizations, and even Pope Leo are trying to forge the appropriate governance path for AI.
AI is just another tool; but, in this case, we need to set boundaries on its use, the open framework and the validity of data. The problem is that AI is really a people pleaser; it wants you to be happy, so it will converge on an answer you want with multiple queries. What if you don’t know what you want?
In many forums, when I have been asked about the impact AI will have on our profession, I simply do this exercise: I ask the audience to raise their hands if they’re a project engineer, project/program manager or principal. Usually that’s about half the audience. Then I say you can lower your hand if you have had one workday where you had a list of things to accomplish that day and completed it by the end of the day. Few, if any, lower their hands. The reason is simple: there are so many variables in our everyday work that a “wash, rinse and repeat” process doesn’t work. While AI tools can inform the process, if there isn’t a validated history of the exact same issue, same location, with the same labor and the external variables, human reasoning is a critical component.
Finding a Framework
As I’ve been working with ASCE, ACEC, GHD and the State University of New York at Buffalo, there are many issues to consider as we navigate AI. Fundamentally, generative AI isn’t a substitute for careful, thoughtful, well-reasoned critical thinking. AI can produce output with hallucinations that are incorrect and misleading. For example, there are many cases of citations that were made up and don’t exist. So, as the saying goes: “buyer beware.”
AI risk isn’t just a technology challenge, but an organizational governance challenge. Issues include workforce governance and implementation, overall readiness, and sensible risk management.
Earlier in 2026, the ACEC Institute released a report, “Leading Through AI Risk: The Enterprise Framework for Engineering Firm Leaders.” The overview states: “Artificial intelligence is rapidly reshaping engineering practice, creating new opportunities to improve productivity, decision-making, project delivery, and client value. But as AI becomes embedded across engineering firms, it is also introducing risks that extend far beyond the technology itself.”
The report examines the risks AI creates across the engineering enterprise and provides a practical framework to help firms identify, govern and mitigate them responsibly. (For additional insight, an accredited webcast on the ACEC report can be viewed at iimag.link/mHlhj.)
The central finding states, “Artificial intelligence does not reduce the engineering profession’s responsibility. Instead, it raises the standard for governance, engineering judgment and organization leadership. Across stakeholder groups, the research found strong agreement that engineers remain responsible for the quality of the work they deliver, regardless of how AI is used.”
Some of the issues that come to mind include the following:
• Any use of AI should be disclosed along with the verification of sources. Transparency and disclosure are critical.
• What’s the reliability of the data? Are the data and its references correct? Ultimately, the engineer and their employer are responsible and liable for their work. If it doesn’t feel right, don’t use it.
• AI is not a substitute for a professional engineer’s license. Who is ultimately in responsible charge?
• Increasing the use of technology introduces cybersecurity risks that need to be identified and mitigated.
• Early engineers learn their craft by simple and repetitive designs. There’s a risk that AI will minimize or even eliminate those jobs. How do we train the engineers of the future as to what actually works—or if it’s good or bad—if they don’t experience that firsthand? You need to understand “garbage in, means garbage out.” How do we teach judgment?
There are many great insightful comments from individuals across the sector. Jamison Braun, senior vice president and managing director for SAP’s U.S. Public Services Business, said, “AI ambition today is actually moving faster than the operating foundation. Agencies should identify high-volume workflows where a backlog, reconciliation burden supply exception, auditability problem or service delay is visible and measurable. Start small and expand. You’ve got to deliver one measurable outcome, and then you have to scale what works.”
We need to embrace the change, but with a proper risk profile.