Television shows, movies and books have long detailed human-robot interactions, and some of the fantastical images in popular culture decades ago are finally coming to fruition.
Siri, the personal assistant available on newer models of the iPhone, can mimic emotions and hold limited conversations with the user. People also can receive laser eye surgery thanks to an invention developed by a University of Arizona physician. And just recently, Esquire released its "Talk to Esquire" application, one the specializes in voice recognition, enabling users to communicate with some of the magazine's experts.
Given the pervasive nature of technology, and the rapid growth of artificial intelligence, what does the future hold for human-robot interactions?
TOPIO, the "TOSY Ping Pong Playing Robot, " is a bipedal humanoid robot designed to play table tennis against humans. The robot gave a demonstration during the Tokyo International Robot Exhibition in 2009. (Photo credit: Wikimedia Commons)
While at the UA, Raquel Torres Peralta worked with Kobus Barnard, an associate professor in the School of Information: Science, Technology, and Arts, and others to investigate and improve gesture and user recognition in robots.
"We will see in the near future more robots that look like a human, gesticulate, laugh and even joke," said UA alumna Torres Peralta.
"Imagine having an android as a surgeon, as a consultant, as a personal trainer or a security guard, all in one," she said. "In the far future we hope could have machines capable of learning from natural instruction, so you can have a servant – yes, a servant – at your command so you can dedicate your life to what's important, but we are still far from that point."
Torres Peralta, who earned a doctorate in computer science in September 2012, continued working at SISTA through December. Peralta has since begun a postdoctoral position at the University of Sonora in Mexico and took a moment to answer some of our questions about artificial intelligence and what types of connections humans and robots may have in the future.
Q: What are some of the emerging themes in the area of robot intelligence?
A: One emerging and interesting (and challenging) research topic is the generation of human-instructable computing to have machines capable of learning directly from us as any other human does. The human-robot interaction (HRI) field has studied many aspects of our relationship with smart agents. One of them is the analysis of how humans teach a robot in order to construct new algorithms capable of learning from natural instruction instead of complicated direct programming by specially trained engineers.
Q: What are the desired outcomes human-robot interaction researchers expect?
A: If we are capable of decode the intentions behind instructions from human teachers, then we can teach a robot how to "digest" natural language and build knowledge from it. This implies having robots as smart as babies that could learn practically anything we are capable of teaching. Other interesting goal in the HRI field is to improve the way robots respond to humans in order to be accepted and make humans around feel comfortable and confident. This is an important issue if we consider that robots will interact with people on a daily basis in the near future, but it is challenging to recognize the emotions in humans and learn from the reactions and respond accordingly in an assertive way.
Q: Whether it be in computer science and computing, the medical field, education, military defense or other areas and disciplines, why must we focus research on improving robot intelligence? What are some of the important implications of having developed intelligent robots and machines?
A: Leaving aside the economical impact, having robots close to what a human is capable to do (as learning from experience, reason and generalize concepts) could improve the quality of our lives. Imagine having someone to help you with no schedule, with no requirements for food or salary and never gets angry. This makes a huge difference if your health is compromised and you depend on others for the most basic tasks every day.
Q: How close are we to having regular access to highly developed robots?
A: Even when the media shows a big progress in the development of intelligent agents, we are still far from having a machine capable of learning as we do. For a robot, a simple step is hard, since keeping the balance while standing or walking requires a number of operations, and the task is more challenging if the machine carries an object or moves through an irregular terrain. Thus, having a robot that can resemble a human being in all the roles and functions is hard to conceive soon. So far, the robots we have seen can perform dedicated tasks (as folding clothes, play a game), but they have to be trained for that and cannot grow their knowledge to learn other tasks by themselves. This question is not simple, in fact, the implications of having fully developed intelligent robots and machines are still not known, but if we are able to have a robot that can replace a human at any position, then humanity will have to evolve to another phase where robots cannot compete with us.
Q: Given the prevalence of artificial intelligence in popular culture (consider "The Matrix," "The Terminator, 2001: A Space Odyssey" and also in literature like Asimov and Manga), why do humans seem to have almost a natural attraction to evolved robotics? Is it a natural inclination at all?
A: I personally believe that we have a natural inclination not to robots per se, but to everything that implies fantasy or superiority (as superpowers) and the figure of robots has been exploited by literature and even publicity as a superior intelligence beyond ours (or stronger, powerful) and this makes us be more curious about them. In experiments we have performed at our lab we have asked people to teach a little robot that looks like a toy about a specific task, and in most cases participants have treated it as a pet or a little child, depending on the robot's performance. When the robot responded as a smart entity the teachers treated him as such. None of our cases had rejected the robot, on the contrary, most people seemed to enjoy the time spent on the session.
Q: Humans seem already to maintain at least some level of close interaction and even interpersonal relationships with devices, such as computers, smartphones and vehicles. How might our emotional and psychological connections to robots/machines change over time?
A: The psychological connection will depend on the way robots are introduced, or how the interaction with them evolve. In the case of smartphones, our relationship with them has come to the point of being seen as an extension of the body, a need or part of the identity, but smartphones don't answer back to us, they are a device with no eyes or face. With the development of androids with very realistic faces and attitude there is a chance of seeing them as other entity at our same level. I would not be surprised to see virtual actors in the movies or online services of boyfriend/girlfriend made by request to cover the emotional needs that some persons cannot cover in the real world. Obviously, the reaction to a robot will be different depending on the situation. We definitely will see some robots in the military in the future, and the fear to that day is not unfounded.