Robots want to be our friends – At least their makers want us to like them
Robots represent a somewhat adventurous representation of the Internet of Things. In principle, they have connected devices that use data to make decisions in real-time and automate processes that are repeatable. They are the essence of the IoT. However, they are frequently designed to be cute and friendly. Even some industrial robots have faces made up of big cartoon eyes on a tablet screen. There is something about humanizing robots that change the relationship dynamic a bit, and the psychological exploration of human-robot interaction has been a growing field, especially as robots are gaining momentum as social companions and personal assistants.
Some initial psychological impressions
An MIT Media Lab report told the story about how a researcher performed an ongoing research project testing out how preschool children would respond to a robot. The robot – a blue stuffed creature with some plastic, robotic features for its face and hand – would sit in a chair with Ethernet cables trailing from it while the researcher would hide nearby with listening to students interacting with the robot and asking the children questions. At this early stage, the robot only had a speaker, microphone and remote controls, so it was in no way autonomous.
However, after asking the robot about their favorite animals, the preschoolers seemed to get attached. They wanted to teach the robot things – even though they were aware the robot was being controlled by a person – and they seemed to somewhat affectionate toward the object. Over years of similar studies and research, the experiment revealed an overarching theme – humans almost invariably built relationships with robots. Individuals did not treat the robots precisely like other humans but would socialize with them. Part of the underlying psychological issue here, according to the report, is that humans are evolutionarily trained to see agency and attention in the interactions around them. Because of this, people tend to identify robots as a thing with intentionality.
Digging a bit deeper into human-robot interaction
Research published by Science Direct explored the way humans and robots interact by delving into how robots respond to dynamic environments that are partially unknown to them. Generally speaking, in human-robot interaction, there is a need for the robot to model human activity and cognition in order to make decisions at a human level. This requires a blend of a variety of artificial intelligent techniques, including symbolic reasoning, task planning, reactive control, and learning. In fact, robots designed for social setting must be programmed around core theories of mind building. Essentially, robots must be designed to handle a wide range of cognitive tasks to effectively respond to human activity in a natural, human-aware way.
Of course, designing robots to respond accurately to human activities also depends on understanding how humans react differently around robots, but based on the types of products coming to the market, there seem to be a few conclusions dominating the industry:
- Product designers believe people want robots that look friendly.
- Robots must be constructed to withstand occasional bumps while they navigate real-world environments.
- Black glass is often needed for faces because the tint will hide the inner workings of the robot and help humans disassociate the robot from the actual technology.
These were a few discussion points brought up in an IEEE Spectrum report describing the Consumer Electronics Show in 2017. At the event, the report’s author noted that many of the social home robots looked similar, and stopped to ask product designers why this seemed to be the case.
With this in mind, let’s take a closer look at three leading social robot solutions, including one that wasn’t highlighted by the IEEE.
The Furhat robot is a social experiment and robot assistant all in one. The robot itself has an extremely human-like face – so much so that it is often pictured wearing human clothes, such as a fur hat – and has been created with the intention of exploring the next generation of computer-human interactions.
There are plenty of ways for people to interface with computers, Furhat gives the computers a voice in the conversation. The system is built with a custom proprietary operating system that has been built with human behavioral principles as its foundation. Major features of the robot include:
- A multi-language SDK that can handle English, Swedish, German, Japanese and Turkish, among other languages. This SDK can support both speech recognition and synthesis.
- A robotic head that can make gestures and otherwise express animations using computer-generated animations.
- A body tracking system with video and depth cameras that allow for face tracking, multimodal processing, and joint-attention mechanisms.
At this stage, Furhat is primarily a research project exploring how computers may interact with humans, but its potential ability to disrupt how we think about human-robot interactions makes it worth the discussion here. Two more mainstream social robotics systems are:
With two glass eyes, a big white nose and a form factor resembling a cartoon character’s head and neck, Jibo is built to bring cuteness to the world of home robotics. In brand documentation, Jibo is described as sincere, spontaneous and curious. The robot has facial recognition software in place so it can tell who people are and bring up information pertaining to their interest. It is designed to actively respond to the people around and serve as a social companion.
With a form factor similar to a Russian Nest Doll, the Kuri social robot is designed with interactivity in mind. As a result, Kuri has capacitive touch sensors so it can “respond to human touch in a truly personal way” and has gestural mechanics designed to help the robot move in natural ways.
There is still a great deal of work going on in the background when it comes to the development of social robots, making it difficult to pin down specific technologies and hardware systems that are gaining momentum in the sector. Much of what is being developed is highly proprietary in nature. One clear trend across the industry, however, is a desire to make social robots incredibly endearing to foster some sort of positive psychological reaction.