Artificial Intelligence seems to be taking over every industry in this day and age. We interviewed the Director of Systems Engineering programs here at Worcester Polytechnic Institute, Don Gelosh, in hopes of getting some insight into AI and the future of Systems Engineering. Don attended the 2024 International Council on Systems Engineering (INCOSE) Symposium in Dublin, Ireland. In this blog, Don discusses his extensive background in systems engineering, his approach to integrating AI in education, and the broader implications of AI in the industry.
1. Tell us about yourself. How long have you been at WPI?
My name is Don Gelosh. I am currently the Director of Systems Engineering programs at WPI. I have over 48 years of experience, including 26 years serving in the US Air Force as an acquisition engineer, and several years in academia, industry and government. I have been at WPI for over 12 years now. Right before coming to WPI, I was the Deputy Director for Workforce Development in the Office of Systems Engineering in the Pentagon.
2. Can you explain what system engineering is?
Sure! According to the definition from the International Council on Systems Engineering, “Systems Engineering is a transdisciplinary and integrative approach to enable the successful realization, use, and retirement of engineered systems, using systems principles and concepts, and scientific, technological, and management methods.”
However, more fundamentally, in my mind, Systems Engineering is the contact sport of engineering! To be successful, Systems Engineers should be out there meeting people, leading teams, and engaging in all aspects of the engineering and management efforts of their organizations. Systems Engineers should be knowledgeable about and involved in all aspects of soliciting and understanding stakeholders’ needs, developing requirements, designing, testing, deploying, and maintaining complex systems. Systems Engineers don’t have to be the experts in everything, but they need to know the experts and how to get them involved. Of all the engineering disciplines, Systems Engineering is the most people-oriented and success depends on excellent skills in technical leadership.
3. What are some arduous manual tasks you as a SE have had to deal with in your research?
Dealing with stakeholders and requirements is a very time-intensive and tedious task. It must be done completely and it must be done correctly, or the project will never succeed. AI tools can help organize all of this stakeholder information. On the other end is the job of verifying the design solution against the requirements and validating the design solution against what the stakeholders really need. I can see AI tools assisting developers with organizing and making sense of this bi-directional traceability.
4. Have you implemented AI on the job, and if so, how has it improved your processes?
As a professor, I have encouraged my students to use AI models and tools as additional resources of information. For example, I’ve encouraged the use of ChatGPT, but I’ve also directed my students to treat the responses as references. That means they have to properly cite the response as a reference and provide their critical assessment of the response. Do they support the response, do they reject the response, or is it somewhere between the two? It’s fine to utilize the AI response and learn from it, but I want to see some critical thinking from my students as well. So far, the students are following my rules and have provided some great insight by using AI.
5. How have you seen the industry as a whole adopt AI?
According to a recent working paper from the National Bureau of Economic Research, early adoption of AI in the U.S. has been somewhat uneven. This paper referenced a November 2023 Census Bureau survey that showed that AI use was highest among the larger companies. The survey also showed varied use among firms in manufacturing, information services, and health care that are using some AI, with firms in construction and retail using much less AI.
The survey also showed that AI adoption is happening in manufacturing hubs in the Midwest as well as some Southern cities as well as the tech hubs in Silicon Valley and the Northeast.
6. What are some areas that AI could improve to better serve the industry?
I believe in general, there are three main areas where AI can improve. These areas are Provide More Data, Improve the Data Quality, and Improve the Algorithm. AI tools are only capable of processing and learning the information that is provided to them. Therefore, you need to provide as much data as possible to the AI tool on a particular topic of interest. However, that data also needs to be high quality – 100% accurate and very relevant to the topic. The old adage of “Garbage in, Garbage out” is very applicable here. In addition, how the AI tool inputs and processes that data and then responds to the inquiries is very important, so you need the best algorithm you can get.
For an AI tool to better serve a particular industry, the tool needs to have access to an abundance of high-quality data about that industry. For example, if you are using an AI tool in the aerospace industry, you should make sure that AI tool has been “trained” on every engineering and management process and procedure they use, including their entire set of systems engineering life cycle management processes. The tool should also be fed with all of their past design solutions. That way, you will have a much better chance of getting a quality response that is worth something.
7. Do you have any concerns for AI in the long term?
I’m concerned that in the long run we may start to depend on AI tools too much and lose sight of critical thinking. I believe it will get much easier to solicit advice and knowledge from AI tools and we will eventually become complacent with the tools and their results. This is a very dangerous area; we must maintain the capability for critical thinking and just plain common sense. Does the response from the AI tool make sense? Has it been validated by other non-AI resources? What does your gut tell you? We can trust AI tools to tell us what they know, but we can’t just accept that information as fact.
8. What do you think ethical use of AI looks like in systems engineering?
I see AI tools and systems as assistants and references in Systems Engineering. What that means is that the SEs can use AI tools, but they must be critical of the responses, what they get in return. Obviously, SEs using the AI responses as their own work is highly unethical and dangerous. But, using the responses as relevant information and insight on the problem at hand is fine, as long as we examine the responses and use our own critical thinking skills to ensure they provide value. We must also ensure that the AI responses do not violate any intellectual property restrictions.
9. Are you excited about the future of AI and SE?
I would rather say that I am cautiously optimistic about the future of AI and SE. I believe AI tools in general will provide a powerful set of capabilities and will enable great advancements in science, healthcare, technology, engineering, and society. All these areas must work together to improve life. I also believe systems engineering can take advantage of these advancements to evolve and improve the processes of systems development, including both the technical competencies as well as the interpersonal competencies required.
Looking for a master’s in Artificial Intelligence or Systems Engineering? WPI offers cutting-edge programs designed to equip you with the skills and knowledge needed to excel in these rapidly evolving fields.