Image credit: Midjourney

Last word: Autonomous

Katharine Flores  • 2023 Fall issue

When people hear this word, they often think of “autonomous driving,” the exciting and somewhat scary idea of letting AI take the wheel and direct a car to its destination while passengers relax and focus on other things. However, there are many other domains where letting AI control the decision-making process can be advantageous and may produce better results than relying on a human-driven process, without the obvious safety risks of self-driving cars. One of these domains is engineering design, particularly the design of new materials.

AI may take the wheel, but we will still set the destination.

A common misconception is that materials are immutable black boxes, and that only a limited palette is available for any given application. In reality, the 118 elements on the periodic table can be combined in countless ways to make materials, some with functionalities we have yet to imagine. Thousands of years of human history have only scratched the surface of what is possible. Just in the world of metals, there are hundreds of billions of alloys waiting to be explored. AI, incorporating physics-based models as well as human intuition, offers the opportunity to not only make predictions about which new materials we should pursue in this vast design space, but to autonomously select, produce, validate and optimize a material with the complex set of properties required for an application, much more efficiently than humans have done alone.

The role of engineers will necessarily evolve as autonomous design becomes a reality. Nonetheless, our primary tasks will continue to be envisioning the future we want to create and asking the right questions to get us there.

Back to Engineering Momentum