Key points:
- Generative design uses AI as part of an exploratory design process
- Students can learn to work with AI, and not in fear of it
- See related article: 3 critical considerations for generative AI on campus
Thanks to ChatGPT, the whole world seems to be wondering whether artificial intelligence (AI) is here to help or harm us.
While educators may be in deep debate about its risks and promises, we know for certain that generative AI like ChatGPT will transform education and the workforce of the future. In fact, according to a recent BestColleges survey, 43 percent of students have experimented with it—some to cheat, others to simply save time.
Naturally, this is worrisome for educators. How can a tool that hand-feeds answers to a student possibly teach them about critical thinking and problem-solving? Or nurture their creativity? We tell our students to get their facts right and check their sources. But generative AI isn’t always accurate and doesn’t show its work—at least, not yet.
And despite some concerns that have emerged about the use of AI in education, I remain optimistic. Through technology called generative design, we’ve incorporated AI into several mechanical engineering courses at Howard University, unlocking a new level of possibilities for our students.
I learned about generative design in 2018, shortly after Autodesk made it available to industry and education users as a feature within Fusion 360. As an early adopter of this technology, Howard University became one of the first universities to incorporate generative design in mechanical engineering curriculum. Here’s a simple example of why it gives me hope for how certain applications of AI can enhance learning, not take away from it.
Hundreds of ways to make a chair, brought to you by AI
Popular AI tools like ChatGPT can give you an answer quickly, but it may end up being the wrong answer. That’s an engineer’s nightmare.
Generative design, on the other hand, harnesses the power of AI as just one step in an exploration process to produce designs that are based on constraints named by a designer or engineer. Importantly, generative design tools need humans to feed them a lot of information first. They need collaborators.
You can’t tell a generative design tool “make a chair” and expect it to design one you can sit on. You need to tell the tool which kind of materials you’re using, how much weight the chair must hold, how it will be manufactured, its geometry, how safe you want it to be, and how much you want it to cost.
Once you’ve provided enough parameters, a generative design tool will give you something no human could: hundreds of design options within minutes. A decade ago, this would have sounded impossible.
At the end of the process, humans step back in to choose the best options to tweak and refine into final designs. Generative design results are precise, repeatable, inspectable, and validated. In other words, they’re an engineer’s dream.
Once I learned about this AI-powered technology, I knew I had to bring it into the classroom.
At Howard, we are using generative design in our research and have changed our curriculum to expose students to it. Our students use sophisticated generative design features within tools like Autodesk Fusion 360 starting in their first year. By their third and fourth years, they’re ready to work on high-performance vehicle designs and drone research. By graduation, they’re industry-ready with skills in an emerging technology highly coveted by employers in our field.
And our students see that value, too. Student surveys showed overwhelmingly positive feedback for generative design and appreciation for how the technology broadened their minds to countless “what if” scenarios during the design process. The surveys also indicate that generative design increased students’ critical thinking in design considerations and overall ability for engineering design.
At Howard, we’ve seen firsthand how generative design indirectly exposes students to the power of AI, teaches them critical thinking, and lets their creativity flourish. It trains students to tackle the real-world challenges they’ll face upon graduation. So, when they show up to work on their first day as a practicing engineer, they’re primed to design the lighter, stronger, more sustainable, and more reliable products of the future.
So, while the debate continues on the risks and promises of AI, reflecting on what we’ve been able to accomplish with generative design gives me hope for our future with this expansive technology. With the right approach, we can prepare humans to work alongside AI, not in competition.
Dr. Nadir Yilmaz is a professor and chair of the Mechanical Engineering Department at Howard University.