For years, Artificial Intelligence (AI), the development of computer systems that mimic human intelligence, was in the province of science fiction, hyping the imagination of futurists and visionaries.
No more. Tools like DALL-E and ChatGPT have jumped from the realm of imagination to the real world, performing tasks similar to the human brain, with tech evangelists already preaching the changes these technologies will bring in promising industry after industry.
One of them is Jonathan Girroir, technology evangelist at Tech Soft 3D. He provides valuable insights into how AI is a game-changer for the engineering industry.
What kind of impact will AI have on the engineering space? Will designers and engineers see significant changes in their world and daily workflows?
Design Assistance Today
AI has already been making inroads into the design space for several years, of course, via generative design functionality. With generative design, users tell their software program what they’re looking to create and give it certain constraints and parameters to work within. In return, it produces dozens of different design options that meet those criteria.
For instance: Do you need to make a bike frame that weighs no more than 5 pounds, is only made of plastic and can withstand up to 300 pounds? Generative design has got you covered and will provide scores of design possibilities.
We will see more and more of this type of role for AI, where it serves as an assistant embedded within computer-aided design (CAD) programs, working side-by-side with the designer and springing into action.
AI can also start carrying some of the load when putting the finishing touches on a design. For example, by typing in a prompt to “make this building look appealing,” a designer can trigger AI to populate an architectural drawing with 3D models of just the correct type of furniture, perhaps, or some perfectly manicured trees and hedges out front. All the designer has to do, meanwhile, is approve AI’s suggested additions.
Streamlined Simulation Tomorrow
If design assistance is the first big area where AI will impact the engineering world, simulation is the second.
This is simply a logical progression of the ability of AI to learn from large data sets and make sense of them in ways that a human couldn’t. It’s easy to imagine, for instance, that AI could predict the result of a very complex computer-aided engineering (CAE) operation — like the effect an earthquake would have on the structural integrity of a building or that a head-on collision would have on a vehicle — with nearly perfect accuracy.
Ultimately, AI won’t be able to build a house or engineer a car. Still, it will be able to assist significantly in the process, taking more and more items off of humans’ plates — significantly increasing the productivity of individual designers and engineers in the process.
New Ways of Creating Digitizations
AI also has some exciting implications for reconstructing and digitizing the physical world. Case in point: We’re now at a stage where we can feed AI a couple of basic 2D images of a particular building or object, and it will create a full 3D, volumetric representation (i.e., not just a superficial surface model) of that item.
The thanks here go to neural radiance fields (also known as NeRF). These AI-powered rendering models ingest multiple 2D viewpoints and then perform internal calculations and extrapolations to generate a 3D model from those 2D images.
That’s not all: There is exciting research around taking this approach even further with AI-bolstered signed distance fields, which encompass multiple aspects of image rendering. Soon, we’ll be able to use marked distance fields to capture not just the geometry of a 2D image but also the light and the materials — all from a couple of pictures.
Don’t Forget Point Clouds
Of course, as photos and 2D images become increasingly viable source material for creating a 3D model, point clouds — created by scanning objects or structures — will remain a valuable source of data.
AI has some tremendous potential applications here as well. Similar to the way AI has become quite adept at feature recognition in photos — identifying the furry, four-legged thing in a picture as a “dog” while identifying the rectangular object as a “couch” — it can bring similar capabilities to point cloud data, helping to pick out hyper-specific features within the sea of data points.
Amid all these AI-assisted developments, however, there are implications for the graphics capabilities of engineering software. For example, many products already manage both mesh and point clouds. But soon, they may have to work signed distance fields, NeRFs and other representations, all while finding a way for the different models to coexist.
Of course, that’s part of the beauty of game-changing innovations like AI. As its impact ripples throughout the larger ecosystem, other technologies must respond in kind to the new world it creates, spurring even more innovation.
Nothing to Fear
After a slow build, AI has reached a tipping point — and the engineering world will feel its impact in ways that range from extensions of what is already possible to surprising new capabilities and functionality. But, for the designer or engineer, this is nothing to fear.
While AI might be the number one game-changer in the coming years, it promises to change the game interestingly, helping designers and engineers tackle their work and shape the world around us with greater efficiency, creativity and new levels of virtuosity.
AFP
