To:
Cynthia Breazeal, Personal Robotics
Pattie Maes, Fluid Interfaces

When I first applied to the MIT Media Lab in 2012, I was a wide-eyed recent graduate looking for some way, any way, to continue the multidisciplinary pursuit of a dual major in mechanical engineering and art. I did not know what I wanted to research, just that I wanted to be surrounded by the experts and uniquely fascinating projects at the Media Lab. Applying today, six years later, I come with an entirely different and specific goal: to dive deeply into the psychological power of companionship interactions, and the incredible accrual of new user interfaces and data on the behaviors of everyday life.

In 2014, I was accepted into a program that would, as cliche as it sounds, change my life. The Google Creative Lab had just started an experimental founders program called 30 Weeks, and I was invited as one of 16 inaugural designers and engineers from around the world to participate. The goal? Go from zero to startup in, fittingly, thirty weeks.

It was during this program that I developed MOTI, a smart companion to help users build better daily habits from exercise to flossing to mindfulness. He was a small, minimal tabletop social robot that utilized lights, sounds, and haptics to convey motivational states, powered through a wifi connection and back-end smarts.

The comical thing is, I hadn’t set out to build a social robot. My original intent was to solve the problem of motivating good habit formation, and to do so using the process I knew best: human-centered design. This process led from the foundational work of user research and scientific literature sweeps, to building a number of crude prototypes of mobile apps, wearables, and physical products (see portfolio). User testing showed a clear and unexpected winner: a small, squishy koala bear with an arduino and a push button that made a happy noise when a user pressed it, indicating they had completed the habit. One user described the experience, smiling and saying: “it’s like you’re impressing someone, even though it’s not actually a someone.”

This was the moment that I discovered the sheer power of companionship tech. Over three years, 2500 user touchpoints, and hundreds of prototypes later, the same insight stood: MOTI was able to elicit a pseudo-social response that tapped into our innate desire for social gratification (and social pressure) for performing behaviors. By being an emotional interaction rather than a transactional one, he could develop a relationship over time with users, gaining their trust and holding them accountable akin to a human (and sometimes better than one!).

The fact that our brains work this way utterly fascinates me. Perhaps more fascinating: it held true when I started applying the same principles from MOTI to conversational AI products in the digital realm versus just the physical. As our technological ecosystem grows and accommodates more of these social and emotional user experiences as standard (read: products like Amazon Echo, Google Duplex, or Anki’s Vector), we have an increased onus to apply them to enable people to become their best selves. Similar to what we’ve seen with social media, there are good ways to use this power, and bad. My goal at MIT is to develop a deep expertise in the nuanced levers that drive this sort of social and psychological connection with technology.

However, this is just one piece of the puzzle. I am similarly fascinated by the proliferation of the number of user interfaces and data points we have on everyday behaviors. When I imagined the future visions of MOTI, it permeated into the many I/O points of daily life: your current location and stress levels to best time interventions, cross-referencing wearable health data to showcase progress, utilizing your smart speakers, lights, and TV screens as triggers, etc. A sort of “Jarvis” for behavior change. When building Luna, a conversational AI for sleep behaviors, I designed her to be an agent that could jump between text messaging, voice, and even Chrome extension messages. We are quickly moving into a post-mobile world; the technology of tomorrow works not just in your pocket, or within a single device, but permeates across the entire set of your daily locations and interactions.

Accordingly, the consumer products of the next decade and beyond will only be as powerful as the data sets that they sit upon. Whether it be predictive algorithms to determine the best time and place for a trigger, the analysis of hundreds of thousands of data points on sleep, exercise, and productivity for coaching insights, or the deep learning behind a dynamic and custom-fit personality of an physical agent, I want to keenly understand the profundity of artificial intelligence on new user experiences. One of the most illuminating parts of assisting a few venture capital firms last year in the due diligence of social robotics startups was how I was often paired with an expert in artificial intelligence to complement my understanding of robotic UI. My second goal at MIT is to develop the knowledge around this new frontier needed to build complex social-emotional tech interactions.

What I am doing now has very little to do with either of these two macro goals - but it was only in straying from my passions that I realized how deep they are. I decided to take my current role as Head of Product at an EV Charging company earlier this year for two reasons: one, to understand how a growth stage startup operates from the inside, and two, to learn how to be the best manager of both products and people that I can be. My future goals include founding another company, and these skills will no doubt assist in that ambition. However, no matter how hard I try, I cannot stay away from lectures and meetups about robotics, conversational AI, and novel UIs. I crave to become a thought leader in these elements of the future.

I’m lucky enough to consider Andrew Ng as a career confidant, and he once told me something that I will always remember: there is no substitute to learning under someone who is an expert. For a number of years, I have been flying by the seat of my pants, learning by doing, and taking roles completely out of my comfort zone; I was 23 when I was Professor Matheus, 24 when I was the first person hired into SC Johnson’s design group without a master’s degree or an accredited design degree, 25 when I started my first company, and 28 when hired to lead the product department for a $100 Million company. However, I’m ready to switch my approach; I agree that the best way to develop an expertise is by being surrounded by mentors and peers who are experts themselves.

The best place to find those experts is without parallel the research being done in the Personal Robotics and Fluid Interfaces labs. (I should note, that the Affective Computing lab is also on this list, despite not accepting applicants. I hope to be able to engage with them for advice and cross-functional project work). In the Personal Robotics group, applying social robotics to the elder care space is of interest to me (what’s the next Paro, that’s more approachable than Elli-Q?), as well as any in the making related to the Media Lab’s Advancing Wellbeing Initiative. In the Fluid Interfaces lab, projects around sleep and attention such as Dormio, AttentiveU, and Masca, have a clear behavioral goal while using novel types of physical interfaces and data points. I would be ecstatic to build off my existing work with MOTI and Luna in either of these groups (though ideally jointly), or apply myself towards a new endeavor intersecting with developing lab theses and my personal goals. I am planning on being in the Boston area soon (where I am coincidentally from originally), and would love to discuss where this intersection lies with Professors Breazeal and Maes. While I intend on a Master’s Degree, I am open to PhD work if it makes the most sense for the course of my research.

My hope is that my skill sets from both industry and academia will be an asset to my research and peers. I am no stranger to academic research, having spent years as an undergraduate researcher, including publishing a paper with the Yale Grab Lab. I am also intimately familiar with user research and the principles of human-centered design, which I believe can only help academia further its impact on end-users. My background in combining art, design, and engineering makes this process overarchingly more holistic, while my ability to prototype in both hardware and certain elements of software enables the rapid cycles of building and testing necessary for the creation of new ideas. I would be honored to be a part of the place that is at the forefront of the technology and user experiences of the decades to come.

Portfolio Link: https://www.kaylamatheus.com/mit-portfolio