Artficial Intelligence / Big Data / Deep Learning + Product Design is:
- The future, no doubt!
- Smart if it enhances the designer’s ability instead of trying to replace him. Otherwise bound to fail.
- Pretty stupid. There are bigger problems.
- Never going to happen (from a technical side).
- Other (please elaborate in thread)
Okay so this project is still kind of abstract, so please bear with me.
I’m currently working on a side project called AI Design Lab. The website can be found here: ai-design-lab.com
It basically researches the use of artificial intelligence and deep learning in the design process to make the best decisions for any design problem. So it basically runs in the background and learns from the designer’s behavior, analyzes it, compares it to a couple of databases and then suggests changes and iterations. Or copied from the website:
“You provide the inspiration, we provide the intelligence: While you work as usual, our app analyzes your designs in the background and compares them towards over 400 different data points - in real time. It uses insights from fields such as engineering and material science, marketing or environmental impacts. It even tracks past revision cycles to predict your client’s reaction to a design proposal. (…)”
It is honestly very far away from the beta phase, so please excuse the lack of visuals at this point. I would actually like to use this thread to start a discussion about artificial intelligence in product design in general.
Do you guys think an algorithm can ever replace a designer?
Does it maybe make our job easier or would it stand in the way of (human) inspiration?
Is it possible for a piece of code to understand a user better than a designer does? A designer is also a fellow user himself and does more than just quantifying needs. Is that an advantage or a disadvantage?
I would love to hear what you guys think!
On my commute this morning I was listening to Studio 360, where the topic of discussion was “Are computers creative?”
It explores many facets of creativity, but I like the idea that a designer’s job is pretty safe. We create something that never existed before, using the context of our current technologies and culture to push the envelope, no? For a computer to do the same, it would need to be programmed to somehow interpret the current state of the environment available, and produce new and engaging concepts based on that. However, I don’t think it’s much of a stretch to go from deep learning to design / NPD.
While I’m interested to see what an app running in the background can tell me about my process, I feel there’s too much that happens offline and in-person for it to be of much use. It probably wouldn’t know that I was changing a boss location or fillet due to a conversation I had with an EE, but it could probably manage to tell me if I was creating geometry that wouldn’t be acceptable to a certain material or process. Better yet, it would be nice to offload some cognitive processes of things I don’t have the capacity to sit down and organize, such as which CMs would be best for the geometries and intended materials in mind, or possible configurations of similar components.
I imagine something along the lines of “Hey, did you know that the rib right there will create sink marks if you use the PC/ABS blend you’ve spec’d? How about moving it over here, or using this other blend of plastic?”
There’s a lot of potential for AI as an aide, in almost any daily task of design, not just CAD. I do think we are a ways away from an automated design process though.
The aforementioned podcast goes into a little detail about Stephen Thaler and his “Imagination Engines, Inc.” which helped him create over 2000 toothbrush concepts for what became the Oral-B CrossAction. I did a little Googling, but couldn’t turn up any of the design concepts created - that would be really interesting to see!