- Email: [email protected]
How will robots integrate into our world?
Current Biology
Magazine Feature
How will robots integrate into our world? Robotic devices, from the dumb mechanical ones through to artificial intelligence outsmarting humans, are spreading around the globe and into all parts of our lives. How will they fit in with living things, from ecology to society? And could they become a danger for the natural organisms and people they may replace? Michael Gross reports. The Science Museum in London, UK, currently hosts a diverse crowd of around 100 very special guests from near and far. Harry from Japan plays the trumpet; Baxter from the USA is good at tidying up lots of different things scattered around him; Nao from Japan looks really cute and brought the little ones who look just like him; Zeno from the USA can mimic the facial expressions of people looking at him — very popular with young visitors; Robothespian from the UK can recite Shakespeare; Kodomoroid from Japan looks like a real woman but isn’t; at least Kaspar from the University of Hertfordshire, UK, is honest about what he is. All of them are robots — some are experimental prototypes, others commercially available. Robots and artificial intelligence have arrived, not just in the Science Museum, where the exhibition Robots: The 500-Year Quest to Make Machines Human runs until September 3rd, but also in everyday life. Last year, a computer beat the leading grandmaster of the ancient and extremely complex game Go. In January it emerged that several insurance companies in Japan are replacing staff members with IBM’s Watson computer, which was developed to answer questions expressed in natural language and became famous for winning the US quiz show Jeopardy! in 2011. The industrial revolution was all about replacing human muscles with machines. What we’re beginning to experience now is a revolution in which machines replace human workers wholesale, including both their muscles and their minds. And at the same time, the robot revolution may also replace certain animals. We are basically beginning to create new life forms, which means that we’ll have to think about where they fit in both in the ecological balance already heavily damaged by our earlier more
primitive technologies, and in our societies, which have a track record of distributing the spoils of progress unfairly. Moreover, all the philosophical questions around robot ethics and the dangers to human identity in a world populated by robots, which have been a staple of science fiction for decades, are now becoming very real issues that have to be addressed urgently. Origin of new species While the Science Museum focused attention on humanoid robots, there are also developments underway that may add a few new species to the animal kingdom. These may not be immediately recognisable as machines — unlike their distant technological ancestors such as Grey Walter’s robotic turtle (Curr. Biol. (2013) 23, R821–R823).
For instance, some researchers are aiming to combine microfluidics and hydrogel materials with electronics and mechanical actuation in order to develop soft robots that appear like animals. The groups of Jennifer Lewis and Robert Wood at Harvard University, USA, presented a prototype soft robot in an octopus shape that they called octobot (Nature (2016) 536, 451–455). This artificial cephalopod is controlled by microfluidic logic, meaning that fluid transports autonomously drive the catalytic decomposition of the hydrogen peroxide fuel to produce water and oxygen gas and the volume expansion that drives its movements. This enables the octobot to produce mechanical movement without any solid parts or external energy supply. Excess oxygen gas is then vented through microscopic pores. A 3D-printing technique incorporating a variety of materials enabled the teams to optimise the technology rapidly, producing and testing 30 different designs and almost 300 individual robots. So far, the octobot doesn’t perform any sophisticated functions apart from moving autonomously without any hard parts. However, the authors see it as a beginning of an entirely new adventure, stating that: “The octobot is a minimal system
New coworker: The workplace robot Baxter from Rethink Robotics, on display here at the Science Museum London, can learn to carry out a wide variety of tasks and is safe for use next to human colleagues. (Photo: Michael Gross.)
Current Biology 27, R199–R217, March 20, 2017 R199
Current Biology
Magazine ones like the honeybee, it is the right response to replace animals with robots, particularly as the environmental problems decimating useful insects are also having ecological knock-on effects on other animals, including birds (Curr. Biol. (2014) 24, R717–R720).
Soft ware: The soft robot octobot can move autonomously based on microfluidics and the catalytic decay of hydrogen peroxide fuel. (Photo: Lori K. Sanders, Ryan L. Truby and Michael Wehner, Harvard University.)
designed to demonstrate our integrated design and fabrication strategy, which may serve as a foundation for a new generation of completely soft, autonomous robots.” Mass-produced octobots equipped with specific functionalities could conceivably serve a wide range of functions, including camouflaged wildlife observations in the style of the current UK television series Spy in the Wild, continuing a growing trend for the use of advanced technology in ecology and conservation research (Curr. Biol. (2014) 24, R629–R632). Commercial uses could also be imagined, and some sinister minds are probably already working on weaponising the new technology. By contrast, other researchers are working from the desired biological function backwards, developing materials and machines to fulfil it. Thus, the group of Eijiro Miyako at the Nanomaterials Research Institute in Tsukuba, Japan set out to develop an artificial alternative to insect pollinators and achieved that aim by combining new materials with existing drone and information technology (Chem (2017) 2, 224–239). Specifically, the researchers developed a novel synthetic ionic liquid gel (ILG) material that can remain in a sticky and wet state for long periods of time regardless of the weather conditions. Furthermore, R200
the superwettability and electrostatic properties of the material were optimised to match those of insect wings and achieve similar adhesive effects on pollen. The Japanese team applied the new material first to the ant species Formica japonica and showed that the ILG-modified ants could successfully collect pollen from and pollinate garden tulips (Tulipa gesneriana). In the next step towards robotic pollination, they equipped a radiowave-controlled drone with vertically aligned animal hairs coated in the gel material. Using this device, they achieved robotic pollination of Ligustrum japonicum (Japanese privet) flowers — selected as a model because they are commercially available in Japan all year round. The researchers explicitly cite the concerns about bee declines as motivation for their research, stating that their work “should lead to the development of robotic pollinators and help counter the problems caused by the declining honeybee populations.” There is little doubt that with current technologies it should be relatively straightforward to make the pollinating drones independent of the radio controller, letting artificial intelligence navigate between plants and decide where to pick up pollen and where to deposit it. The bigger question is whether, faced with the impending loss of species including domesticated
Current Biology 27, R199–R217, March 20, 2017
Replacing people Being replaced by a robot is also a fate that many human workers are facing in their jobs. Mechanical and low-skilled tasks have long been a target of robot takeovers, but now the combination of artificial intelligence and the robotic replication of many human abilities from face recognition to walking have led us to the point where not only physical work but also mental work of human employees can be handed over to machines. By recent estimates, this may affect up to half of the currently existing jobs, from customer service through to accountancy. Japanese companies, faced with a shrinking workforce, are already embracing the opportunities, replacing accountants with artificial intelligence and customer service personnel with humanoid robots. It is often predicted that the care for Japan’s growing elderly population will soon only be possible with the help of care robots. Elsewhere, the corporate quest to optimise productivity is the main driver behind the robot takeover. Rethink Robotics in Boston, USA, the company that produces the versatile robot Baxter that can be trained on a wide variety of tasks, openly praises the machine as a replacement for humans: “Companies globally have integrated Baxter into their workforce, and gained a competitive advantage.” To entice employers, the website states: “If you walk the floor of your facility and see monotonous or dangerous tasks, then Baxter is ready to get to work for your company.” Baxter is safe to work next to humans, if any remain on the factory floor, and is sensitive to forces, so it can feel if it applies the right amount of pressure for a given task. The number of jobs lost to automation is often cited as a key concern, but on the other hand one can argue that these jobs were often unattractive in the first place, and revolutionary technical developments have so far always created new jobs while destroying old ones.
Current Biology
Magazine The bigger question looming behind all this is how the wealth that robots create will be distributed. In the extreme scenario, when most of the goods and services are produced by robots rather than employees, the very small number of people owning these robots will be the only ones to earn any money from the production. Without a salaried workforce, on the other hand, there will be no customers to buy the goods, and thus the whole market economy will cease to function. Thus, if salaries paid for productive work are no longer the major fluid driving the mills of capitalism, how can societies avoid the extreme imbalance that, apart from being cruel and unfair, would also be likely to kill off the system? The old idea of the guaranteed minimal income paid out by the state, giving everybody enough to survive regardless of their work status, has been revived in various, sometimes surprising quarters. In the current French presidential election, the topic is officially on the agenda, supported by candidates on the left of the spectrum, including the official Socialist Party (PS) candidate Benoît Hamon, but derided by those on the right wing. A region in Finland is currently testing such a scheme and it is being discussed in political circles around the world. Other approaches would involve regarding robotic workers like people and making them pay tax, as Bill Gates suggested recently in an interview. As income tax is related to a salary, this idea may need some refining — what would robots do with a salary? While the populist wave in the USA and the UK is threatening to return society to the 1950s, politicians are left with too little time to consider what life will be like in a world full of robots, due to arrive well before 2050. The European parliament, at least, has recognised the need for some kind of rules and regulations, which are currently being debated. A report recently passed by the European parliament’s legal affairs committee proposes setting up a European agency for robotics and artificial intelligence, legal definitions and obligatory registration for the most life-like robots, ethical codes for the production and use of robots, a reporting structure for companies, and an insurance system.
Drone delivery: A novel ionic liquid gel enabled researchers in Japan to develop a robotic pollinator. (Image: Eijiro Miyako.)
Obviously, a political zeitgeist steering exactly in the opposite direction, will fight any attempt to introduce new regulations, so there is a real risk that the technology will become a danger to lives and society before any reasonable political measures will have been found. Global danger Science fiction has taught us to fear armies of Daleks and Cyborgs, but the bigger danger may lurk elsewhere, perhaps even in your car or your refrigerator. As the cyber security expert Bruce Schneier warned in a recent feature in New York magazine, we are in the process of “building a world-size robot” and we are not even trying to keep it safe. What Schneier refers to is the “internet of things” where all kinds of appliances from home entertainment through to transport are online. Historically, Schneier explains, two different security cultures have co-existed in separate worlds. Aircraft engineers, for instance, work on the premise that their products must not fail, because lives will be lost if they do. Computer and software producers have evolved the rough and ready approach where bugs and security vulnerabilities will occur frequently and be patched up quickly. And the small computers that typically hide in appliances are even less
secure and aren’t getting the benefit of regular patches. The trouble is, as Schneier explains, that the two worlds are now colliding, with the consequence that “Computer security is now everything security.” Self-driving cars (Curr. Biol. (2015) 25, R255–R258), hospital management systems, or infrastructure control centres that are vulnerable to malicious bot attacks or even some new kind of viral program can very quickly lead to large numbers of casualties. And once there are millions of robots replacing the human and possibly insect workforce, they will become part of the global super-robot too, and thus part of the rapidly growing security problem. Like the European parliament, Schneier calls for new regulation to match the challenges of the new technologies. He argues that previous introductions of breakthrough technologies, such as broadcasting, nuclear energy or air traffic, also led to the establishment of state authorities to control them. Given the speed of progress and spread, this now appears to be an urgent task to take control of the development before it overwhelms us along with society and the natural world. Michael Gross is a science writer based at Oxford. He can be contacted via his web page at www.michaelgross.co.uk
Current Biology 27, R199–R217, March 20, 2017 R201
Magazine Feature
How will robots integrate into our world? Robotic devices, from the dumb mechanical ones through to artificial intelligence outsmarting humans, are spreading around the globe and into all parts of our lives. How will they fit in with living things, from ecology to society? And could they become a danger for the natural organisms and people they may replace? Michael Gross reports. The Science Museum in London, UK, currently hosts a diverse crowd of around 100 very special guests from near and far. Harry from Japan plays the trumpet; Baxter from the USA is good at tidying up lots of different things scattered around him; Nao from Japan looks really cute and brought the little ones who look just like him; Zeno from the USA can mimic the facial expressions of people looking at him — very popular with young visitors; Robothespian from the UK can recite Shakespeare; Kodomoroid from Japan looks like a real woman but isn’t; at least Kaspar from the University of Hertfordshire, UK, is honest about what he is. All of them are robots — some are experimental prototypes, others commercially available. Robots and artificial intelligence have arrived, not just in the Science Museum, where the exhibition Robots: The 500-Year Quest to Make Machines Human runs until September 3rd, but also in everyday life. Last year, a computer beat the leading grandmaster of the ancient and extremely complex game Go. In January it emerged that several insurance companies in Japan are replacing staff members with IBM’s Watson computer, which was developed to answer questions expressed in natural language and became famous for winning the US quiz show Jeopardy! in 2011. The industrial revolution was all about replacing human muscles with machines. What we’re beginning to experience now is a revolution in which machines replace human workers wholesale, including both their muscles and their minds. And at the same time, the robot revolution may also replace certain animals. We are basically beginning to create new life forms, which means that we’ll have to think about where they fit in both in the ecological balance already heavily damaged by our earlier more
primitive technologies, and in our societies, which have a track record of distributing the spoils of progress unfairly. Moreover, all the philosophical questions around robot ethics and the dangers to human identity in a world populated by robots, which have been a staple of science fiction for decades, are now becoming very real issues that have to be addressed urgently. Origin of new species While the Science Museum focused attention on humanoid robots, there are also developments underway that may add a few new species to the animal kingdom. These may not be immediately recognisable as machines — unlike their distant technological ancestors such as Grey Walter’s robotic turtle (Curr. Biol. (2013) 23, R821–R823).
For instance, some researchers are aiming to combine microfluidics and hydrogel materials with electronics and mechanical actuation in order to develop soft robots that appear like animals. The groups of Jennifer Lewis and Robert Wood at Harvard University, USA, presented a prototype soft robot in an octopus shape that they called octobot (Nature (2016) 536, 451–455). This artificial cephalopod is controlled by microfluidic logic, meaning that fluid transports autonomously drive the catalytic decomposition of the hydrogen peroxide fuel to produce water and oxygen gas and the volume expansion that drives its movements. This enables the octobot to produce mechanical movement without any solid parts or external energy supply. Excess oxygen gas is then vented through microscopic pores. A 3D-printing technique incorporating a variety of materials enabled the teams to optimise the technology rapidly, producing and testing 30 different designs and almost 300 individual robots. So far, the octobot doesn’t perform any sophisticated functions apart from moving autonomously without any hard parts. However, the authors see it as a beginning of an entirely new adventure, stating that: “The octobot is a minimal system
New coworker: The workplace robot Baxter from Rethink Robotics, on display here at the Science Museum London, can learn to carry out a wide variety of tasks and is safe for use next to human colleagues. (Photo: Michael Gross.)
Current Biology 27, R199–R217, March 20, 2017 R199
Current Biology
Magazine ones like the honeybee, it is the right response to replace animals with robots, particularly as the environmental problems decimating useful insects are also having ecological knock-on effects on other animals, including birds (Curr. Biol. (2014) 24, R717–R720).
Soft ware: The soft robot octobot can move autonomously based on microfluidics and the catalytic decay of hydrogen peroxide fuel. (Photo: Lori K. Sanders, Ryan L. Truby and Michael Wehner, Harvard University.)
designed to demonstrate our integrated design and fabrication strategy, which may serve as a foundation for a new generation of completely soft, autonomous robots.” Mass-produced octobots equipped with specific functionalities could conceivably serve a wide range of functions, including camouflaged wildlife observations in the style of the current UK television series Spy in the Wild, continuing a growing trend for the use of advanced technology in ecology and conservation research (Curr. Biol. (2014) 24, R629–R632). Commercial uses could also be imagined, and some sinister minds are probably already working on weaponising the new technology. By contrast, other researchers are working from the desired biological function backwards, developing materials and machines to fulfil it. Thus, the group of Eijiro Miyako at the Nanomaterials Research Institute in Tsukuba, Japan set out to develop an artificial alternative to insect pollinators and achieved that aim by combining new materials with existing drone and information technology (Chem (2017) 2, 224–239). Specifically, the researchers developed a novel synthetic ionic liquid gel (ILG) material that can remain in a sticky and wet state for long periods of time regardless of the weather conditions. Furthermore, R200
the superwettability and electrostatic properties of the material were optimised to match those of insect wings and achieve similar adhesive effects on pollen. The Japanese team applied the new material first to the ant species Formica japonica and showed that the ILG-modified ants could successfully collect pollen from and pollinate garden tulips (Tulipa gesneriana). In the next step towards robotic pollination, they equipped a radiowave-controlled drone with vertically aligned animal hairs coated in the gel material. Using this device, they achieved robotic pollination of Ligustrum japonicum (Japanese privet) flowers — selected as a model because they are commercially available in Japan all year round. The researchers explicitly cite the concerns about bee declines as motivation for their research, stating that their work “should lead to the development of robotic pollinators and help counter the problems caused by the declining honeybee populations.” There is little doubt that with current technologies it should be relatively straightforward to make the pollinating drones independent of the radio controller, letting artificial intelligence navigate between plants and decide where to pick up pollen and where to deposit it. The bigger question is whether, faced with the impending loss of species including domesticated
Current Biology 27, R199–R217, March 20, 2017
Replacing people Being replaced by a robot is also a fate that many human workers are facing in their jobs. Mechanical and low-skilled tasks have long been a target of robot takeovers, but now the combination of artificial intelligence and the robotic replication of many human abilities from face recognition to walking have led us to the point where not only physical work but also mental work of human employees can be handed over to machines. By recent estimates, this may affect up to half of the currently existing jobs, from customer service through to accountancy. Japanese companies, faced with a shrinking workforce, are already embracing the opportunities, replacing accountants with artificial intelligence and customer service personnel with humanoid robots. It is often predicted that the care for Japan’s growing elderly population will soon only be possible with the help of care robots. Elsewhere, the corporate quest to optimise productivity is the main driver behind the robot takeover. Rethink Robotics in Boston, USA, the company that produces the versatile robot Baxter that can be trained on a wide variety of tasks, openly praises the machine as a replacement for humans: “Companies globally have integrated Baxter into their workforce, and gained a competitive advantage.” To entice employers, the website states: “If you walk the floor of your facility and see monotonous or dangerous tasks, then Baxter is ready to get to work for your company.” Baxter is safe to work next to humans, if any remain on the factory floor, and is sensitive to forces, so it can feel if it applies the right amount of pressure for a given task. The number of jobs lost to automation is often cited as a key concern, but on the other hand one can argue that these jobs were often unattractive in the first place, and revolutionary technical developments have so far always created new jobs while destroying old ones.
Current Biology
Magazine The bigger question looming behind all this is how the wealth that robots create will be distributed. In the extreme scenario, when most of the goods and services are produced by robots rather than employees, the very small number of people owning these robots will be the only ones to earn any money from the production. Without a salaried workforce, on the other hand, there will be no customers to buy the goods, and thus the whole market economy will cease to function. Thus, if salaries paid for productive work are no longer the major fluid driving the mills of capitalism, how can societies avoid the extreme imbalance that, apart from being cruel and unfair, would also be likely to kill off the system? The old idea of the guaranteed minimal income paid out by the state, giving everybody enough to survive regardless of their work status, has been revived in various, sometimes surprising quarters. In the current French presidential election, the topic is officially on the agenda, supported by candidates on the left of the spectrum, including the official Socialist Party (PS) candidate Benoît Hamon, but derided by those on the right wing. A region in Finland is currently testing such a scheme and it is being discussed in political circles around the world. Other approaches would involve regarding robotic workers like people and making them pay tax, as Bill Gates suggested recently in an interview. As income tax is related to a salary, this idea may need some refining — what would robots do with a salary? While the populist wave in the USA and the UK is threatening to return society to the 1950s, politicians are left with too little time to consider what life will be like in a world full of robots, due to arrive well before 2050. The European parliament, at least, has recognised the need for some kind of rules and regulations, which are currently being debated. A report recently passed by the European parliament’s legal affairs committee proposes setting up a European agency for robotics and artificial intelligence, legal definitions and obligatory registration for the most life-like robots, ethical codes for the production and use of robots, a reporting structure for companies, and an insurance system.
Drone delivery: A novel ionic liquid gel enabled researchers in Japan to develop a robotic pollinator. (Image: Eijiro Miyako.)
Obviously, a political zeitgeist steering exactly in the opposite direction, will fight any attempt to introduce new regulations, so there is a real risk that the technology will become a danger to lives and society before any reasonable political measures will have been found. Global danger Science fiction has taught us to fear armies of Daleks and Cyborgs, but the bigger danger may lurk elsewhere, perhaps even in your car or your refrigerator. As the cyber security expert Bruce Schneier warned in a recent feature in New York magazine, we are in the process of “building a world-size robot” and we are not even trying to keep it safe. What Schneier refers to is the “internet of things” where all kinds of appliances from home entertainment through to transport are online. Historically, Schneier explains, two different security cultures have co-existed in separate worlds. Aircraft engineers, for instance, work on the premise that their products must not fail, because lives will be lost if they do. Computer and software producers have evolved the rough and ready approach where bugs and security vulnerabilities will occur frequently and be patched up quickly. And the small computers that typically hide in appliances are even less
secure and aren’t getting the benefit of regular patches. The trouble is, as Schneier explains, that the two worlds are now colliding, with the consequence that “Computer security is now everything security.” Self-driving cars (Curr. Biol. (2015) 25, R255–R258), hospital management systems, or infrastructure control centres that are vulnerable to malicious bot attacks or even some new kind of viral program can very quickly lead to large numbers of casualties. And once there are millions of robots replacing the human and possibly insect workforce, they will become part of the global super-robot too, and thus part of the rapidly growing security problem. Like the European parliament, Schneier calls for new regulation to match the challenges of the new technologies. He argues that previous introductions of breakthrough technologies, such as broadcasting, nuclear energy or air traffic, also led to the establishment of state authorities to control them. Given the speed of progress and spread, this now appears to be an urgent task to take control of the development before it overwhelms us along with society and the natural world. Michael Gross is a science writer based at Oxford. He can be contacted via his web page at www.michaelgross.co.uk
Current Biology 27, R199–R217, March 20, 2017 R201