Humans have long envisioned mechanical servants; references can be found as far back as the fourth century BC, and appear in the writings of ancient Chinese, Muslims, and Leonardo da Vinci, who designed a mechanical knight who could sit up, wave its arms, and move its head and jaw. In the 1700s, automatons, or mechanical humans, could act, draw, fly, or play music (History of Robots, 2017).
Like these automatons, robots in popular culture tend to take on human characteristics. Visitors to the 1939 New York World’s Fair could see Westinghouse’s Elektro, a seven-foot-tall robot that could speak 700 words (enough to say, “I’m smarter than you,” which he did), smoke cigarettes, blow up balloons, and move its head and arms (Blakinger, 2016). Elektro was accompanied by his faithful dog Sparko, who could bark, sit, and beg.
The Star Trek character Data epitomizes the human-appearing robot (and was an evolution of Spock in the original series). A reoccurring theme in Star Trek: The Next Generation was Data’s attempts to become more human, to understand human idiosyncrasies and emotions. In fact, Gene Roddenberry told Brent Spiner, who played Data, that over the course of the series Data would become “more and more like a human until the end of the show when he would be very close, but still not quite there” (Data (Star Trek), 2017).
So that is one way to think of robots, with their ultimate goal to achieve as many human characteristics as possible. We continue to see robots like that, such as Pepper, “the world’s first emotional robot,” according to its creators, SoftBank Corporation, a Japanese telecommunications giant. Pepper, a sort of proto-C3P0, is designed to interact with humans and is advertised as “the premium platform for connecting brands with audiences.” Pepper can identify inflection and tone in human voices and is designed to move naturally.
Some robots are pets. A search for “robot dog” on Amazon generates many pages of hits. In Japan, Sony sold 150,000 of its robot dog AIBO before discontinuing it (AIBO stands for Artificial Intelligence Robot and sounds like the word pal or partner in Japanese). Boston Hospital is testing a therapeutic robot teddy bear called Huggable.
Yet, as impressive as they may be, robots such as Pepper will never be mistaken for people, and your robot dog is little like a real dog (although AIBO owners may disagree). If you assess how common robots are in our society and your definition of a robot is a machine that attempts to assume human or animal characteristics, then robots are a niche product yet to greatly influence society at large.
Another Way to Look at It
Yet, if we consider a robot to be a machine capable of performing tasks traditionally performed by humans, then robots are everywhere. Robots work in Amazon’s warehouses, bringing items to stockroom workers, who no longer need to roam the warehouse to retrieve inventory items. Robot bartenders mix drinks on Royal Caribbean’s cruise ship Quantum of the Seas. Robot pharmacists fill prescriptions. Police robots are sent into hazardous situations, defuse bombs, and have even been used to go after barricaded suspects. Roomba vacuums clean houses. Modern factories are filled with industrial robots. Cars are increasingly autonomous. The military extensively uses drone weapons. The list goes on.
If we broaden our definition of robots to include virtual entities, robots become even more common. Virtual robots are often known as bots. Web crawlers are bots that automatically troll the web, visiting millions of sites and collecting information. Googlebot, for example, creates the vast index that Google uses when responding to search queries. Other web crawlers are more nefarious, impersonating people by leaving comments on blogs and forums or creating user accounts (fake Twitter accounts, for example). Chatbots can converse with people, perhaps on dating sites or to provide technical support. WeChat users, for example, can converse with a chatbot to hail a taxi, order food, buy movie tickets, book a doctor’s appointment, and so on. In the financial world, bots (also known as automated trading systems) buy and sell stocks, sometimes initiating huge transactions within milliseconds, based on market conditions, breaking news, or tweets. This list also goes on.
An Increasing Factor
The fact is that robots are increasingly a factor in our world, and while physical robots are far from being human, virtual robots are becoming increasingly human, and are passing a kind of Turing test – for example, when a person responds angrily to an online comment posted by a trolling bot, he has mistaken the bot for a person.
While there are a large number of benefits associated with these technologies, there are also risks as the prevalence of robots raises a host of concerns. Chatbots, for example, should behave responsibly and provide accurate information. This is a more manageable issue if a chatbot is, say, accepting a pizza order, than if a chatbot is giving medical or financial advice. How can chatbot developers ensure that their responses are correct, and how can their correctness be verified? If the chatbot is programmed to learn, how can developers validate that learning does not introduce errors? Should the chatbot require acceptance of terms and conditions before proceeding? Do chatbots track data? How can a threshold be set to determine how well the chatbot understands a question? At what point does the chatbot request human intervention (if at all)? Can the chatbot recognize that it is talking to a child and respond accordingly? Can the chatbot recognize if it is talking to another chatbot?
These issues are upon us. In 2015, Swiss police arrested a bot. Performance artists Carmen Weisskopf and Domagoj Smoljo had created the bot, named Random Darknet Shopper, gave it $100 in bitcoin to spend per week, and turned it loose on a shopping spree on the darknet, the illicit side of the internet. Random Darknet Shopper purchases included counterfeit sports shoes, fireworks, and ecstasy pills. These items were displayed at an art gallery, with the ecstasy purchase leading to the bot’s “arrest” (Peitzker, 2016).
The police eventually realized that all had been done in the name of art and released the bot (the code, actually). I believe we will see more and more instances of robots breaking the law, intentionally or, increasingly, unintentionally, as robotic technologies continue to advance and become more prevalent. In the next article, I’ll take a closer look at what may happen in the future and how we can deal with it.
Blakinger, K. (2016, April 30). A look back at some of the coolest attractions at the 1939 World’s Fair. Retrieved from New York Daily News: http://www.nydailynews.com/new-york/queens/back-attractions-1939-world-fair-article-1.2619155
Peitzker, T. (2016, September 5). The first chatbot arrest, but what are the implications? Retrieved February 19, 2017, from VB: http://venturebeat.com/2016/09/05/this-is-the-first-chatbot-to-be-arrested/
Wikipedia. (2017, 02 15). Data (Star Trek). Retrieved from Wikipedia: https://en.wikipedia.org/wiki/Data_(Star_Trek)
Wikipedia. (2017, February 11). Elektro. Retrieved from wikipedia.org: https://en.wikipedia.org/wiki/Elektro
Wikipedia. (2017, February 11). History of Robots. Retrieved from wikipedia.org: https://en.wikipedia.org/wiki/History_of_robots
Dr. Craig Rosenberg is an entrepreneur, human factors engineer, computer scientist, and expert witness. You can learn more about Craig and his expert witness consulting business at www.ui.expert.