Miku Hirano is an AI scientist, tech-savvy entrepreneur, young mother, and co-founder-CEO of Cinnamon, Inc. in Japan, one of the Forbes “25 Machine Learning Startups to Watch in 2019.” Motivated to solve problems with enabling technologies, she is a prime example of a moonshot leader. GLOBIS Deputy Dean Dr. Jorge Calvo sat down to discuss her work and vision for the future.
A Moonshot is Born
JC: Do you consider yourself a moonshot thinker?
MH: I’m not sure how closely I fit the definition of “moonshot,” but my personal goal is to move humankind forward. If you think about human history and primitive eras, first we found fire. Then we made tools to help us make food. Later, we invented cars, trains, and the computer. Now we have the internet and smartphones.
I think of this as irreversible technology. Take smartphones, for example. Before smartphones, it was quite normal for us to not have access to the internet all the time. But now, if we leave our phones at home, we feel completely lost. That dependency is irreversible, and it’s the kind of technology I want to make—the kind that moves humankind forward.
JC: Does this passion come from your family?
MH: Yes, from my family. Specifically, from my first son. Around when I had him, I heard some really sad news about a young Japanese lady, about 24 years old, who committed suicide because she was working too much. Unfortunately, this wasn’t an isolated incident. Other people in Japan have committed suicide due to overwork (what we call karoshi in Japanese).
Currently, especially in Japan but really everywhere, people work too much. It feels normal, but it really shouldn’t. After my son was born, I started to think, “This isn’t the kind of work environment we should leave for the next generation.”
JC: Can you tell us a little about how you transitioned into entrepreneurship from university?
MH: My major at Ochanomizu University and degree at Tokyo University Graduate School were both in computer science, and for most students in that field, a master’s in science is pretty normal, so I followed that path. I also did some internships during my bachelor’s—at Goldman Sachs, Panasonic, Recruit, those kinds of companies—and I found I wasn’t really good at being an employee. So, after graduation, I went for my master’s degree at the same time as I founded my first company.
JC: How did you form your vision for Cinnamon?
MH: I’ve always hated boring, repetitive tasks, and that has a lot to do with our vision. I was very inspired by a book I read that said focusing negative feelings is the key to starting a new business or service. Basically, needs are based on negative feelings. When you feel, “This is so annoying,” that’s a business opportunity.
JC: That’s actually one of the core concepts of moonshot thinking—take something really painful, or something you really hate, and change it.
MH: Exactly. So that’s what we’re doing at Cinnamon.
AI in Japan
JC: What is the state of AI in Japan?
MH: Well first, you have to understand that there are 4 AI fields: business AI, internet AI, factory AI, and autonomous AI. Business AI is for office workers, handling things like call centers and documentation. Internet AI is like Uber or TikTok, which use machine learning to provide better user experiences. Factory AI is, for example, Amazon, which uses robots for storage. Autonomous AI is self-driving cars, that kind of thing.
There’s a common misconception that Japan’s AI (and IT) is really behind, compared to the US or China. And yes, three of the four fields are really behind in Japan, but not business AI. Japan’s business AI is really number one. Even compared to China and the US.
Business AI is the field Cinnamon is in. We opened our sales front in the US so we could compare with our American competitors, and we’ve seen that Japan is a full year ahead compared to the US. They’re doing things now that we were doing a year ago in Japan.
In China, we couldn’t even find any competitors. Investors there are only interested in self-drive and those kinds of “sexy” AI projects. Other “unsexy” AI companies cannot get the Chinese to invest.
JC: Cinnamon is your second startup. What motivated you to start a second AI company?
MH: My first startup (acquired by MIXI) really wanted to be a global company, but we didn’t make it. That was my biggest regret in my first startup, so I decided to try again. For Cinnamon, I decided to found it outside of Japan to make the global leap easier.
JC: So, going global was part of your vision from the start?
MH: Yes, our vision is to become the number one global Business AI company. Our ideal goal is to remove all the repetitive and boring tasks from the world. That’s the challenge we set for ourselves. It’s still a ways off, since we still only have four products. To increase our product lineup and technology, we’re hiring AI researchers very aggressively.
JC: What kind of disruption does Cinnamon provide?
MH: Most of our clients are banks and insurance companies. They use business process outsourcing (BPO) quite often because they need to do lots of data input with 100% accuracy. That means they often use two or three people to do data entry for just one document.
What we’re doing is replacing one of those people with AI. When companies lack labor, employees need to work longer. AI can remove 30–50% of that burden, and it can run for 24 hours a day. Cost differentiation and speed are also values we provide. We have a lot of AI talent in our people, and we’re committed to understanding our clients’ needs. We start talking with clients even before launching products.
JC: Where do you see Cinnamon in 5 years?
MH: We’ll be the number one Business AI company in the world.
Pushing AI to the Future and Beyond
JC: Can you tell us a little about what Cinnamon’s AI does and where you hope it will go?
MH: Our main products are Cinnamon AI, which is a document reader, and Rossa Voice, which is a speech text engine. Both are based on recognition, but humans are more intelligent. We can do more than just recognize. For example, if I read a document you send me, I can recognize that you want to emphasize this or that. Currently, AI doesn’t have that intuition, but we hope to get there someday.
JC: There have been some predictions of when AI will be able to write a book, or even replace AI researchers. Do you think the hype about AI has impacted expectations?
MH: I think what we can do with AI right now is much lower than what people expect—not just for us at Cinnamon, but for any AI researcher.
Having said that, AI doesn’t develop the way the human brain does. If we’re talking about a novel, actually yeah, the current technology can handle that because there’s no correct answer for a novel. If you want to write a report, though, even at the university student level, that’s very difficult for AI.
JC: How far do you think AI will develop in the next five years?
MH: The way I see it, there are four phases to innovation. It starts with geeks, then big corporations, then small-to-medium enterprises (SMEs), then individuals. For example, if you think about websites in the 90s, only geeks were really excited about making websites. Then big corporations like Dentsu started spending millions of dollars to make their sites. After that, SMEs got onboard. And now, even individuals can have their own site for free.
Currently, AI is in the second stage: large enterprises. In 5 years, business-wise, I’m guessing it’ll go to SMEs. What’s holding it back is that currently, AI technology cannot provide general solutions. All AI companies need to customize their projects. However, if we can accumulate the right technologies or operations, we can make AI scalable for the SME phase.
JC: Do you have any ethical concerns about AI?
MH: There are a lot of negative things in the media about AI, but a lot of it is click bait. They say AI is going to replace human work—those kinds of things. But the truth is, all technologies replace human work.
One hundred years ago, people needed fire to cook, but now we can push a button. Fifty years ago in Japan, train stations needed a person to stand at each gate and take tickets. Now we have ticket machines and passes. We look at those old practices and see that it doesn’t need to be that way anymore. Humans today are working too much, and we at Cinnamon believe it doesn’t need to be that way.