For over 100 years humankind has fantasied about robots that look, behave, and interact like us. This has been the theme of many science fiction movies from the silent film Metropolis in 1927 to recent feature films like Terminator, Ex-Machina, or the TV series Westworld. However, a group of scientists from Sweden is working on making this a reality.
The start-up Furhat Robotics is developing a social robot that displays complex facial expressions without the mechanical limitations of other technologies, due to its unique back-projected solution. The team has developed the world’s first Operating System for social robots and a dedicated Software Development Kit. This allows developers to customise pre-built expressions and gestures to fit a particular character’s personality.
Innovation Origins interviewed co-founder and CEO Samer Al Moubayed. He spoke about the start-up’s goals and the potential for social robots in daily life.
How did you come up with the idea for Furhat Robotics?
Well, we all come from the Royal Institute of Technology in Stockholm, and we have dedicated our careers to trying to understand human-like machines that can interact with people the same way that we humans interact with each other. A lot of the technologies behind it have come quite far and have since converged. So, several years ago we realized that the time had come to start experimenting a bit more with that. That’s why we combined a lot of our research into a single product, which is the Furhat Robot.
What sets your robot apart from other robots?
The term robot is undefined at the moment. The word describes all types of things from an autonomous vacuum cleaner to a roadside camera that measures traffic flows and even software which are accounting tools. Therefore, people tend to use the word to describe any kind of smart technology that acts in response to something that happens in the world.
But what we are working on are social robots. As in that image in people’s imaginations of something very human-like that is capable of recognizing and understanding our language, our bodies, our facial expressions, and are also capable of speech. Perhaps the closest to this on the market are those AIs like Siri or Alexa. Yet these are still essentially transactional robots. You ask a question and you get an answer. We really want to build a machine that can engage with people on a deeper social level. Very few companies are working on this right now. We’re very different from companies that say they are building robots, we are actually building a social robot.
What is the potential for social robots in the real world?
Our long-term vision is that the development of robots will happen in the same way as it does with any other devices like smartphones or personal computers. What Farhat wants is to not only build one or two applications – but to have thousands of applications. Just like a smartphone has. We’re aiming to create a social robot that has all these applications which can be used across multiple industries. As people start getting used to having social robots around them, more applications can be added. In that respect, the sky is the limit.
But in the short term, there is already a great product on hand that we can deliver for certain circumstances. Social training s a major area, where a robot can simulate social situations. It can be an immigrant, a patient, or an angry customer. People can interact with it so that they can train themselves on how to act in these social situations. Because you can program the robot to have distinctive personalities, i.e., it can be angry, bored, etc. This is an area that is very unique to these robots.
We also have some really exciting projects in Germany where we are utilizing robots at airports and train stations. They can speak many languages and provide information about travel tickets or schedules for any traveler. We’re aiming that in future the robot will be a lot more sensitive and be able to detect any stress or your anger levels and can help much more effectively than a touch screen.
Plus we also have an interesting partnership with a recruitment company which has to do with providing a robot that can perform job interviews. This gives someone the opportunity to have social interaction, just without another human being present. Because humans are very biased and we are influenced by how people look, dress, or speak. This is one way we could eliminate any unconscious bias. We call this unbiased recruitment.
How does the robot work?
We have a substantial set of tools on offer for people who want to develop applications and/or customized data for creating a personality for a robot. But we also have an operating system that runs on our robot, which is the world’s first operating system for social robots. This is comparable to what Windows is for your desktop, where you have apps running but Windows is the operating system that connects everything together. Our operating system does the same, except with language, cameras, and microphones that all run on the robot. All our algorithms are available as open-source for developers. This means that they can use this data to install applications on a robot. We also have technology that can change the face of the robot, so you can have a child robot, or a woman, or a man, or faces with diverse ethnicities. This is very easy to alter.
And what was your biggest challenge?
We are still in the early days so there will still be a ton of challenges ahead. Nevertheless, whenever we present the robot to society and encourage companies to build applications for it, they often tend to think the robot can do everything. And if it doesn’t, then they want it to do something very specific. Which they then expect can be done within a month because they think it should be easy. But the reality is completely different. Some applications will require tens of thousands of hours to build along with millions of dollars.
But the biggest problem is that companies still do not believe in or see the value of this technology. As a consequence, they aren’t willing to invest that much money. They invest only a small amount, but that doesn’t mean that the robot will be able to handle the application well. It’s a bit like building apps for a phone. Some might just take a couple of days to build. But these apps are not the ones that will become billion-dollar apps and change the world. People don’t seem to have that same understanding when it comes to social robots. Therefore, our biggest challenge is usually figuring out how to motivate companies across the various markets to not only scratch the surface but genuinely invest in the long term.
You can read more IO articles about start-ups here.