December 11, 2018 by Luke James
While many people believe the advent of artificial intelligence is going to steal away all hard-working professionals' jobs, that is actually the opposite of the truth. AI enhances engineering and helps engineers streamline and perfect their designs more than ever before.
There has been plenty of talk regarding how AI could “kill†engineering jobs. This is simply not true.
Whilst AI will change the nature of engineering and the profession itself overall—and yes, this may involve the “killing off†of certain jobs and skillsets currently fulfilled by humans—AI's impact will provide new opportunities for the most talented engineering professionals to adapt and innovate.
This is a discussion for another time, however—there is plenty of literature surrounding this dotted across the web.
Engineering is expensive. It is iterative. It eats up a lot of time. Human engineers play a huge part of this; they come up with innovations and designs based on their own knowledge and experience, which culminates in an end result that is then tested and optimized. This is great, but human engineers are only capable of coming up with a finite, usually quite small and specialized, number of ideas or solutions for any given problem due to our limited brain capacity, processing power, and, ultimately, lifespan.
Although this has not been a problem throughout the industrial revolution and the early onset of computing and technology, it is this relatively inefficient and limited R&D process that causes the biggest shortfalls. It is where we are beginning to realize the drastic improvements that AI could make.
With AI, R&D teams can take a data-driven approach and augment their own intelligence to come up with new ideas or supplement their own, instead of solely relying on their own areas of expertise and specialist knowledge. Utilizing machine learning and data, AI technologies will help the next generation of design engineers to draw up ideas quickly and efficiently, bringing together elements from other industries and areas where a human engineer may never have looked.
This is known as generative design.
Through the process of generative design, everything from computers and cars to tables and kitchen utensils is being reimagined. With a limitless scope, AI can come up with designs that are beyond the minds and imaginations of human engineers, delivering solutions and products for industries and consumers that would otherwise never have been developed.
Utilizing AI software and computing power offered by the cloud, generative design enables design engineers to solve virtually any design-related problem simply by providing an existing design—or by describing and defining it through simple parameters such as materials, budget limitations, height, and weight. With this information, the generative design software backed by AI can, in some cases, come up with thousands of different options, far more than an entire team of human engineers could come up with.
Image courtesy of Autodesk.
It does this by leveraging machine learning to draw inspiration from the evolutionary design approach. Exploring all possible options from the given design parameters, the results outputted by generative design software can be analyzed and filtered down by human engineers to select the one that is most suitable for a given project.
Instead of starting the CAD process by physically drawing up a potential design from their own memories, experiences, and ideas, design engineers can simply tell a computer what they are looking for or what they want — or describe a problem that they are trying to find a solution for.
If, for example, you are designing a circuit, instead of spending valuable time drawing up two, three, four, or more potential designs, you can simply tell a computer that you want a circuit design that uses X material(s), costs Y to produce, and has Z number of a specific component(s). Instantly, you will potentially, given your parameters, see hundreds, if not thousands, of possible design options that meet your criteria. On your own, this would be impossible.
Seeing generative design in action perhaps provides the most compelling insight into its true power. Imagine the perfect track race car that performs flawlessly with the perfect size, weight, aerodynamics, and strength.
Image courtesy of 3ders.org.
This design feat would be impossible for humans alone. However, Bandito Brothers managed to pull this off using a generative design-based approach that used over 4 billion data points collected from sensors attached to a traditional racing car. With this data—alongside instructions fed to it by the Hackrod team: “design the lightest possible frame relative to structural integrityâ€â€”the Hackrod design was born a few hours later.
Given the huge problem space—the number of ways a car can be designed is virtually limitless—this required immense computing power. However, this project's purpose was to represent the limitless potential of generative design; typical design production does not have a problem space this immense.
Generative design has even been applied to creating an office floorplan.
Image courtesy of BIM+.
When planning for a move to a new office in Toronto, Autodesk surveyed 250 employees who would be working in the office.
Questions probed what the employees valued in their workspace—natural light, proximity to colleagues, and outdoor views, for example—and then the results were quantified before being written into an algorithm used for generating the office’s layout.
This algorithm used this data to design and optimize an office layout that would satisfy the desires and quell the pain points associated with the employees who would be working there.
Although this approach did not save much design time and could not factor in costs considerations, it demonstrates that generative design has a huge scope that will only grow as the tech behind it improves.
In engineering, there is a generational gap and it is perhaps more prevalent here than in any other. This assertion is highlighted by the fact that many young engineers are now using AI to the extent where it forms a core part of their workflows whereas ‘older’ engineers are not.
This is to be expected, though. Engineers who are just now entering the scene in their twenties are digital natives. They were people born just prior to and at the beginning of the digital boom. These people have grown up with tech: they rely on search engines to find information, their interest in engineering was likely piqued by using consumer tech products at a young age, and, overall, they understand it better.
The fact of the matter is that AI is already widely being used by the world’s youngest and most talented engineers–these people can appreciate the benefits that implementing AI has not just to their workflows, but to their designs too.
AI’s growth is inevitable, not just for engineering but for virtually every other industry. Getting ahead now whilst it is still emerging rather than having to play catch-up in future is the best move you can make now, especially if you have just come to the sad realization that you are in fact an ‘older’ engineer (even if you are only in your forties!) who thought AI’s scope is limited to the likes of smart home assistants.
A future where AI is heavily involved in engineering, particularly design engineering, is far from a bad thing.
At the moment, design potential is being held back by an outdated iterative process that consumes far too much time and does not explore new avenues. Liberating design engineers from these constraints, AI's application to engineering in the form of solutions such as generative design will allow them to take on higher-level tasks and truly apply their talents to the more valuable project and workflow elements.
