Design Thinking Applied to Machine Learning Systems
We are in the midst of massive change. Right now many entrepreneurs, engineers, and researchers are busy rethinking or redesigning some of the big sectors such as travel, business, remote work, collaboration to name a few. With dependable advancement in Machine Learned (ML) features and tools in our hands, it has become a great opportunity to rebuild new experiences that could have a long-lasting impact on our economy and society.
10 years back Tim Brown (then CEO, IDEO) gave a compelling Ted talk on moving from Design to “Design Thinking”. I have watched it several times. Recently, it occurred to me that several points made in his talk are still very relevant. When we think about a new AI-first product — we usually think of something a bit more attractive (smarter ways to figure out something from context?) and making it a bit easier to use?
But according to Tim, in Design Thinking the focus should be less on object/product/feature and more on Design thinking as an end-to-end approach. Instead of taking the existing convergent approach of selecting the best choice among existing solutions, it allows us to explore new alternatives that have not existed before. This way we might actually see a bigger impact in the long run.
According to him, the following are the three main pillars of Design Thinking as an approach:
Human-Centered
It may integrate technology and economic components but it starts with asking human-centric questions such as:
What humans might need? What makes our life easier, and more enjoyable?
What makes technology (AI) useful, or usable?
Do we understand enough about culture and context?
Do we understand the motivation and aspiration of users?
How other humans might play a role in helping the real, needful users?
As an example, think of an AI-based healthcare system being designed without considering the above questions?
Learning by Making
Faster prototyping is the key to understanding the strengths and weaknesses of our ideas. The faster we build prototypes, the faster our ideas evolve!
Instead of thinking about what to build, building in order to think.
I see many tech companies nowadays have adopted this idea in order to get early feedback on new products/experiences.
Participation as Main Objective
Instead of seeing the primary objective of products as consumption (by end-users), we should explore the potential of active engagement of everyone in meaningful experiences.
Many forms of value beyond simple cash could be created and measured
The greatest impact happens when the product is designed with participation from everyone. For example, Wikipedia and other products based on collaboration.
In summary, I really see great value in the above ideas when considering a new design/product (with or without AI). Machine learning is just a tool but we need to keep our focus on participation and the human-centric aspect of any future systems.
Disclaimer: Above thoughts are mine and it has nothing to do with my employers.
