Written by Olivia Walter. The original post can be found on Medium.

Machine learning and user experience design…do those two phrases even go together?

How I Found Out About Machine Learning and UX

When I decided to get into tech, I had several years under my belt of leading personal transformation and spirituality events. These varied in size from small and intimate to large, multi-day seminars. I have also worked with children, and been a professional musician. In short, I had way more experience with people than with technical skills.

In the spring of 2017, I started my journey of tech exploration. As a people-nerd, but not a tech-nerd, I’d never heard the words “machine learning” until my partner described it to me, who was just starting a job at an AI company. I began researching the topic and found myself watching YouTube tutorials, going to Meetups and events, and generally obsessing over this thing called machine learning. Around the same time, I discovered UX (User Experience design) and decided it was an intuitive entry point into tech. I can work with people AND technology? Score.

I was into my first few weeks at GrowthX Academy (a 3 month long intensive that trains tech talent outside of coding), learning the ins and outs of UX design when I found myself wondering — if there such thing as UX for machine learning?

From what I could tell, the people who were training neural networks and mining data were not firstly concerned with the experience of their users. And the designers who were gaining empathy for their users were not even aware of how data was being used to make programs smarter.

There seemed to be a pretty big gap there. Then, I thought about the billions of internet users around the world. How many of them knew what was happening with technology? I saw an even larger gap.

How important is it for users to understand the technology they are using? After all, a painter need not have an intimate understanding of the design, manufacturing and shipping process of the paintbrush they use. They just need to use it for their own purposes. The knowledge gap between maker and user is going to be a given. But when the paintbrush is capturing every motion and move the painter makes and uploading it to a database that then creates algorithms based on that data to make recommendations for more paintbrushes on completely separate platforms…? Ok, this isn’t happening with paintbrushes. But it sure is happening with internet applications.

I decided to do a little research. GrowthX has an amazing curriculum that is designed to teach UX design and research with real applications. So, I used a class research assignment to interview over a dozen people about their understanding of machine learning. I found out that of the half that was not already developers or tech people, they had no idea how the technology worked. Moreover, they found many aspects of this new technology to be “creepy” and “strange” but very “convenient and useful”.

Would people want to know that the information they put into their browser or smartphone feeds into machine learning algorithms? Would they want to know the mechanics of it? Would they even care? Would it make them more discerning about what information they do share?

What I found while studying machine learning is that I would need to be a mathematician and programmer to fully understand the basic concepts. What if people who are curious don’t even have a chance to get past the barriers to entry to learn about this stuff? We are ignorant of even which things we are ignorant about!

As I pondered this, I decided to find out who else was having similar ruminations. A quick internet search was surprisingly barren but was not completely dry. The article Machine learning and UX — it doesn’t exist yet was my entry point. Then I discovered a local Meetup that was less than a week away — Let’s MLUX!

After class the next week, I walked over to Quid, where the Meetup was being held and found myself meeting UX designers, data scientists, and engineers. It was as if this event popped up in perfect timing with my curiosity. So clearly there is a connection being made, at least in San Francisco, the hub of technology. Since then I have continued to meet other interested in this intersection. While the topic of machine learning and UX is new, I have a strong sense that there is a rich story yet to unfold.

Going Deeper

User experience is all about empathizing with the people using your product and designing with their experience in mind. It can make the difference between a delighted user and an alienated one. User experience has a place, I believe, in the emerging machine learning industry.

Individuals and organizations need assistance in shifting both practice and mindset when it comes to using machine learning in businesses and products. And at the output layer of those neural nets are real people with real emotions, and many of them have no idea how artificial intelligence works. Many of them get freaked out when Facebook knows the details of their Amazon searches. Sure, a gap may always exist between maker and customer. But no one should be left behind if they truly want to learn. And no one wants to feel alienated by technology intended to assist them in life. There needs to be a way to talk about what is happening with technology as it gets more advanced, in a way that we can all understand, and contribute to.

UX, I believe, has the potential to address this.

As much as I love geeking out about neural networks and learning about the difference between deep learning and applied AI, it may be a long time, if ever, that I learn about it as an engineer. People will always be my specialty. I care deeply about how technology affects humans and one thing I do know how to do is to help bring people together. I can volunteer at events. I can organize Meetups. I can start online discussions and write blog posts like this one. I can reach out to people via public channels and start conversations.

I will continue to seek out opportunities to learn. Even if we are approaching the singularity, I believe that we can keep our humanity — flawed, beautiful, messy, inspiring, and ever be growing.

So where we go from here?

In the past few months of discovery, I have seen, heard, and thought of ways that UX can be applied to machine learning. As I continue to learn about machine learning and the field of AI in general, more posts will come going into more depth.

  1. User interfaces designed with an understanding of what user actions do to change the results of algorithms.
  2. Using the UX research process to discover how to design ML powered platforms that engage with the users according to their own mental models and sense of ease. Let’s face it — many people prefer talking to people over machines, even if they don’t even realize their customer services reps have been bots for more than a few years now. Whether it is a chatbot, recommendations, or voice recognition — empathy is needed to discover what will give the user a sense of ease and safety. So many people freak out at the idea of artificial intelligence. Positive experiences and associations will improve the user experience of AI.
  3. Design user flows that are able to help users understand how machine learning is working to produce the results. Not in a way that only a Ph.D. or programmer would understand. Think a business owner with enough data available that it makes sense to implement machine learning. Or even an everyday internet user who wants to know more but has less refined math skills.
  4. Integrate UX into AI teams to catch cognitive biases that may be going into the datasets in order to prevent things like sexist or racist bots.
  5. Mindset training for business owners, product managers, and employees to make the shift into an increasingly more digitized market.
  6. For those privies to clever visual design…come up with some better stock photo ideas for artificial intelligence!

Donald Knuth has known for the quote: “Computers are good at following instructions, but not at reading your mind.” The computers of today and the future are able to do something that seems eerily similar to that. Our behaviors and actions have effects and consequences. Perhaps machine learning can teach us more about the human experience than anyone could have predicted. As technology grows, let’s grow with it.