Challenges in Human-Robot Interaction: "The Future of Robot Business", an interview with Dong-Soo Kwon

KAIST Professor Dong-Soo Kwon, Ph.D, is probably one of the most well-know robotic experts in Korea: he is the vice-president of the Korea Robotic Society (KROS) and the Director of the Human-Robot Interaction Center at KAIST, where he and his team research about Human-Robot Interaction, Medical Robots and Haptics. Because of his reputation both in academia and in industry, I have decided to ask him what are, in his opinion, the prospective of future robot business: will home robots ever take off as commercial goods?

Let's compare the robot industry to other industries which share similar technologies and purposes, such as the car industry (purpose: mobility), the computer industry (purpose: information technology) and the mobile-phone industry (purpose: communication).

The car industry dates back to the Benz Patent-Motorwagen (1886), but it is only from 1908 with the Ford Model T and the advent of mass production that we can talk of a real car industry. Furthermore, it took until 1972 to reach saturation point (1 car per driver in the U.S.A.) and the point in which the market became completely mature: from the start to the saturation point it took 64 years. The computer and mobile-phone markets followed a similar evolution - introduction, mass production, saturation point. So for instance, the time elapsed between the first mass-produced computer by IBM in 1981 (IBM 5150) and today (we have now finally reached the point in which we have 1 computer per person in the U.S.A.) is about 30 years.
Similarly, the mobile phone market started with the Motorola DynaTAC in 1983 and can now be considered mature, after about 27 years.

So, what about the robotics industry? We could say that robots started being mass produced in 1975 with the PUMA Unimation, which would make robots present on the market for more than 30 years. According to the paradigm followed by similar industries, robots should hence already be a mature market. This is in fact true, but it is only limited to the case of industrial robots: any modern factory uses some kind of production industrial robot, confirming that the market is already mature. However, when we talk about personal robots, we can probably say that only in 2002 with the Roomba, robots made their appearance in the consumers' home as a mass produced good. Taking 2002 as a symbolic date, we can infer that the personal robot market is still very young (about 8 years old) and that if the paradigm of other industries still holds for robots, we should have a mature home-robot market only in 2035 (a common symbolic date in many sci-fi books and movies, including I, Robot).

The next questions to ask are what we will use robots for and how much will they cost. Professor Kwon has no doubts: we will use personal robots in a variety of situations for a multitude of purposes. Personal robotic toys will entertain and babysit children, teachers will use telecontrolled robots to teach in remote places (telepresence), while restaurants and touristic destinations will make use of service robots to better guide and serve customers: in many cases though these robots will not be completely autonomous but simply partially guided remotely by other humans (the service providers), allowing them to serve multiple customers simultaneously. In terms of cost, we can again compare the robotic industry to similar ones and try to drag some conclusions from that data. If the price of an ordinary car is in the $10,000-$100,000 range, the price of a computer in the $500-$3,000 range and of a mobile phone in the $50-$1,000 range, we can probably expect personal robots (which share some technology with each one of these products) between $3,000 and $10,000, a price that should be cheap enough to make robots appealing for consumers and assuring profitability to the industry.

Further readings:

Roti by KAIST and Rastech Co.: English Tutoring Telepresence Robot