Industrial and systems engineering covers many industries, roles, and positions. The main focus of an industrial and systems engineer, or ISE, is continuous improvement. With such a flexible charter, we provide value in any industry in which we find ourselves. However, there are commonalities across industries as we navigate the transition to Industry 4.0, the term for the current industrial revolution in which digitization is driving enormous jumps in innovation. We find ourselves problem-solving for digital thread, or the linkage of an entire product life cycle in an easy-to-access digital record; artificial intelligence, or AI; and logistics.
For a successful outcome, we need a diversity of viewpoints in these areas. Modeling problems requires a variety of paradigms through which to view situations. Each lens will show a different aspect and will reveal observers’ blind spots.
While ISEs are working on these challenges as a discipline, we must beware of different biases that may affect these systems. As we design products and equipment for the 90% of the population that fits within a bell curve, we must study and engage the diverse population who may be outside that curve to ensure their needs are met or face the possibility of staggering unintended consequences.
Digital thread
In my industry, aerospace, a lot of discussion and effort go into the development and management of a digital thread linking data from all stages of the entire life cycle of a product. Think about it: How do you know what part was installed on which line unit? How do you know the requirements that led to that part selection? How do you know where the materials were obtained and how the part was made? Currently, these bits of data are tracked in multiple systems over a given enterprise (and sometimes through several companies). Having a commonality of systems and a common data warehouse will be powerful for improving the efficiency of all steps of the process.
ISEs are central to these conversations because we have the toolkits to develop systems that acknowledge the interconnectedness of multifaceted data. Process mapping is something we have long been associated with, and this new dimension of data architecture has allowed this skill set to shine.
Breaking down what is meant by life cycle reveals the complexity. This includes original product requirements, the design process for the product and the production systems, and the delivery and aftermarket servicing of the product. Commonality of data or data structure is an important foundation as they flow through the design and production systems. Commonality enhances automation because it is easier to repeat processes with input and output data in a standardized format. It also helps enhance analytics, enabling preventive interventions instead of reactionary ones. Streamlining the retrieval of information will also assist with root-cause corrective actions when problems arise and will even assist with audits.
ISEs are already masters of standardization, so it is easy for them to apply these principles to data management and designs of systems. They know how to account for the interconnectedness of multiple functions, making them ideal problem-solvers in this new frontier.
Artificial intelligence
AI is affecting everything. The key right now is understanding how to train it effectively and ethically. ISEs are working to identify training models and results testing methods to ensure the AI is learning properly. ISEs will look at the reduction of variation in the process and the optimization of answers. Their education in quality systems helps with testing models to ensure failure modes are identified and mitigated. ISEs are also key to building process controls to set alerts for when things might be going out of the control of the system
Labor shortages are beginning to slow across some industries, but the need to attract skilled, committed young peopl
Our testing needs to account for the complexity of, quite frankly, everything. AI seeks to make connections to data that exist, but in an effort to make connections, it can make logic leaps that are bad. An example would be lawyers who used ChatGPT for a legal filing and cited cases that did not exist.1 Testing can help predict, then mitigate these outcomes. Know how a system will fail, and you can put in process controls to prevent that from happening. Process failure modes and effects analysis is an example of a tool that can be used in these efforts.
The ethical challenges with AI are vast. Irresponsible development can lead to imbalances with harmful outcomes. ISEs can use root-cause corrective action models to help map out potential ethical pitfalls that can occur because of the training models. Think of an autonomous vehicle: Testing and learning will need to occur for the AI system to know whether, for example, something in the road is a plastic bag or a rock and what the optimum decision chain should be for a best outcome. Braking for a plastic bag might be unnecessary, but braking for a rock could prevent significant damage to the vehicle.
Especially within AI, identifying and accounting for biases, including gender biases, are key to successful development. The woman engineer’s perspective is needed to ensure applications will not have unfavorable biases against women.
Logistics
Since the COVID-19 pandemic, a lot of attention has been drawn to global supply chains and how they affect what we are able to purchase and at what price. ISEs have long worked on supply chain processes and systems to minimize costs and disruptions. However, the complexities of the world introduced new variables into the system, leading us to develop more sophisticated models to adapt.
In a textbook scenario, risk is a finite amount, measurable by examining past performance. For example, a good year for a wheat crop is one with a certain amount of precipitation, distributed over a particular amount of time, with temperatures in a certain band. From history, we can look at the frequency distribution of ideal versus nonideal conditions and make predictions. But now with the climate crisis and new extremes in weather, past data is not as predictive of future risk. Instead, we need to identify new variables and countermeasures to minimize risk and deliver the products consumers need.
Women’s perspectives are key in any product development. Without accounting for the interactions of as many biases as possible, we can have significant unintended consequences that can alienate large sections of the population. That can be devastating as we pursue these fields going forward. Digital thread, AI, and logistics all need diversity to solve the complicated problems facing us today. Women bring the experiences and perspectives that will push this field forward and help make the world better because of our involvement.
Footnote
1 “These lawyers used ChatGPT to save time. They got fired and fined,” Washington Post, Nov. 16, 2023.
