So, you want to be an AI designer

I published a much longer version (~2800 words) for this article on ACM’s <interactions>, grab it here if you are interested.

AI, Artificial Intelligence — One of the hottest and sexiest words in 2016 and 2017 and the year beyond. But this word makes me nervous — it makes me so nervous that I don’t want to say it loud every time I introduce myself to you.

Designer? AI?! Wow!

YOU: So, what do you do?
ME: I am a designer… 
YOU: Cool! Where do you work?
ME: Hmm, I am working at an AI Lab… 
YOU: Did you mean “Artificial Intelligence”? Wow!

I said this word “AI” in such a low key. But you still heard it, clearly. I saw the surprise and curiosity in your eyes, lightning; I saw question marks as well as exclamation points in your mind, exploding; I saw your mouth become wider and bigger in an accelerated motion — “Wow!” And then smaller, along with a speedy wind coming out from the gap between your teeth —

Cool!

This is what makes me nervous.

The Truth

I am not cool

I am cool in a lot of ways but not the way in your mind. I don’t consider myself as an AI designer or a designer designing for AI products which sounds pretty cool. I have to tell you the truth that I am just a normal designer designing products you are familiar with — mobile apps, websites or wearables. I am doing the same things — brainstorming, sketching, pushing pixels or talking a day non-stop with users.

The Myth of AI Products

What is AI? Everyone has its own view. For me, AI is an enabler and it augments human intelligence. AI itself cannot be products. If we put AI as the core and become too attached to it we will start to ignore the mission of the product — to solve a human problem. Every technology is a tool but not the problem; as the creator, we need to focus on the problem and detach ourselves from the tool — it cannot bring value until we find the right problem. I learned the lesson after trying to innovate AI products for almost a year.

But it is a fantastic tool — AI is so fantastic and powerful with so many potential to solve enormous human problems. That is why I am here — to find the right problem and design AI-enabled products in the right way.

I do not know

I do not know AI; I do not know Machine Learning or Deep Learning or Supervised or Un-supervised learning; I do not know Algorithms, Modeling, RNN, CNN, LSTM, Latency, WER; I do know Math — but it is too basic for AI/ML. I do know product and design and psychology. I do know I will not be an AI/ML researcher/engineer but I am a designer — building products with the mission to bring technology into products and into people’s daily activities — to make life a little bit easier and better, from which the value of technology can be truly shown.

You May Ask

How can you do your job when you do not know so many?

“I know a bit and I have taken Stanford’s Machine Learning course on Coursera…” When asked what you know about AI/ML, some people would respond like this. I have to admit that I only watched the first video of that course taught by my big big and big boss. I gave up learning after I found out there is a lot of math in it (kidding…). My goal is not to become an AI/ML expert but to design products which are AI-enabled.

Learning how to build the right product and build the product right is my goal. Of course, knowing something about AI/ML is essential. The “something” is the core of AI — the concept, the history, the potential and the limitations. Anything else, like designing any other products, I will make my efforts to investigate and understand the problem — reading books, trying out different AI softwares and talking to AI/ML researchers/engineers. I always learn the things when they are needed and when there are clear assumptions and goals.

What is the key for designing AI-enabled products since it sounds like learning ML is not necessary to you?

There are tons.

  1. It is the same as designing any other products: Know the problem well, understand user needs, keep business goals in mind, keep stakeholders in the loop, test early and iterate often… You name it.
  2. Find the right problem and a good use case: It is extremely important for AI-enabled products. As I mentioned in the beginning, sometimes, our mind is too focusing on the technology itself that it becomes too narrow and limited to notice the right human problem. Not to mention to have empathy for users.
  3. Ask Why, a lot of Whys: We cannot solve the problem right unless we identify the root cause. Ask Why, a lot of Whys — “Why this user is complaining that our app sucks?” “Oh, he is angry that the transcription is coming out slow. Why is it slow? Is it because of the internet or the region or the device or the way it coming out?” “Why does user feel slow? What could be the right pace?” “Or maybe there is something can be improved for the language model?” Do a lot of research, ask a lot of people — because it utilizes such an exquisite tool — AI, it is so easy to see the problem on the surface but takes time to reveal the root cause.
  4. It can be better and we can do better: Technology can be better but cannot be 100% perfect. Face, Speech, Image Recognition, Natural Language Understanding… The easier for human, the more difficult for machine. We are born with those skills which are engineered for machines — there might be similarities but more differences. When there is a technology limitation, we still can use the power of design to make the experience better — there is always a way. It is not wise to rely on the technology while blame it when bad things happen. We are human, be human, use the human intelligence, think the human way and do the human thing.

Have more questions? Please feel free to ask me :)

So, you want to be an AI designer

I am not an AI designer but a normal designer as everyone else. I cannot hold this privilege or be labeled as “AI Cool“. Everyone is born to serve — either serve themselves or others.

The same technique for answering each question — “Why?” Why are you interested in AI? What aspects of AI attracts you? What is AI? What do you think are the opportunities and challenges for AI? How do you imagine your daily life designing for AI? And a lot of Whys…

Try — try to use products that make good or bad use of AI, try to think about problems that AI can help to solve; talk to people — AI experts, Creators of AI-enabled products, Users who like and are using those products in daily life… Then ask yourself a lot of Whys again until you find the root cause of your “want”.


Andrew Moore, the Dean of Computer Science Department at CMU gave us a talk about AI and its future. My heart was beating fast when listening to his inspiring talk. In the end, I raised my hand and straightened my body — asking:

“Thank you so much for the talk, Andrew! It is very inspiring though I am not an AI researcher/engineer. I am a designer with backgrounds in psychology and HCI. I am always thinking about how we can apply those amazing technology like AI into solving real problems and making people’s life better. For non-AI people, someone like me, how can we help and get involved? Or, what kind of extra help do AI researchers/engineers need to tackle those exciting while challenging problems?”
“In the long history of humankind (and animal kind, too) those who learned to collaborate and improvise most effectively have prevailed.” — Charles Darwin

Recommend

<Superintelligence: Paths, Dangers, Strategies>

If you love my little writing, please recommend! Stay tuned for my next story about designing for AI :)