The field of human robot interaction is a widely expanding field. One of the key topics of human robot interaction is interaction with people. This is different from the research done by strictly algorithmic and engineering robotics work. HRI tends to focus on not just artificial intelligence and engineering, but also cognitive science and psychology.
Assume the robot works. Now what?
So there is a lot of amazing research how to make robots work. The algorithms to help them learn are becoming more and more advanced. The mechanical engineers have developed amazing control systems that are only getting better. But if we assume that everything is perfect from the engineering side, then what? Why do we need robots? What is it supposed to do?
If we look at popular culture, we see TV shows and movies with robots interacting with other people. This has been true for decades. Take a look at Star Wars, 2001: A Space Odyssey, Wall-E, etc. We have done amazing steps toward this Star Wars-esque technology. We actually have successfully populated Mars entirely with robots. We have the technology to put R2D2 and C3PO on Mars, but we don’t actually have the technology to make R2D2 and C3PO.
And that’s fundamentally a design problem. We’ve designed these robots to interact with the rocks and the dirt. Even back on Earth, our robots are still mostly designed to interact with objects and not people. The most common robot in people’s homes is the Roomba, which is literally designed to interact with dirt. That’s the most pervasive robot in the living room.
And in the better part of my lifetime, research has been spent on developing social robots that are designed from the ground up to interact with people. Because in the end, if the robot works, we want them to interact with humans, not objects.
Building a best friend
One of the most profound developments in computing today is artificial intelligence. “Machine learning” is a pretty ubiquitous term, and the more it’s used the less meaning it has and the more it feels like a buzzword. However, there isn’t any denying that it’s not a huge deal. The idea of pattern recognition has been an idea for a few decades in signal processing, computer vision, and other statistical optimization methods.
What’s the most interesting thing is that algorithms are not designed with a computing mindset anymore. We’re shifting to a perspective where we try to devise algorithms that act and think like humans. Computers are becoming less and less mathematical machines. Even though at the core they are, this idea is becoming blurrier and we’re forgetting that they are machines. The same way that we look at other humans and forget that they are just protons, neutron, and electrons. We’re starting to look at computers as a more abstract interface, even if movies like to remind us that they are machines in the end.
Algorithms trying to model human behavior is a big deal. That’s why so much research is spent in devising algorithm’s like the Pragmatic Chaos algorithm which is how Netflix tries to recommend movies to you. Everyone is trying to best figure out how to model human behavior.
With robotics, we want to make that intelligence be more physical. The idea of embodied artificial intelligence is what Rodney Brooks considers true artificial intelligence. Among all the work in developing artificial life and robot life, there are two subdivisions of human robot interaction.
The first is human centered robotics. These are robots designed to help people and assist humans. They exist as aids. To help take care of the elderly. To take care of the ill. To remove humans from factories.
The second is essentially to make a human. I like to refer to it as building my best friend. Building a robot that is not indistinguishable from a human, but is capable of empathy and care. And you forget that inside is wires, silicon, and software. Similar to how you forget that your human best friend is cells, tissue, and nerves.
Am I a robot yet?
There are some interesting questions when we reach that point of more human robots with emotional intelligence. The field of human computer interaction has existed for quite some time, but human robot interaction is a relatively new and more specialized field.
What’s interesting is that humans have become more like computers and cooperative with technology. We are working in conjunction with technology with our phones, laptops, and tablets to form a more productive society than without it. And it’s a very common idea that we’ve become a computer dependent society, and in some ways yes, but we now have amazing ways that our lives are enriched by it. Humans have become to think a bit more like computers as a consequence.
Robots are becoming more and more like human because we want to have machines that we can empathize with. Guy Hoffman gives an interesting talk on devising robots with ‘souls’. What I find most interesting is that while we’re making robots more like humans, robots are also having a psychological and emotionally affective impact on human emotion and empathy.
It seems to me that we’re meeting in a middle where robots are not only becoming more human, but humans are becoming a bit more robotic too.