The robot servant, capable of doing many of the things that we are – cleaning the house, making the dinner, and even bringing us breakfast in bed, then taking away and doing the dishes afterwards: Still seems like a dream right? Maybe only for another 15 years…
Let’s start with the best home robots we already have, JIBO and Pepper. They are not quite able to take the dog for a walk, but certainly point towards things to come.
Described as the first real R2-D2, JIBO looks like a good and even affordable start – $500, available from December 2015. So, from R2-D2, to our own C-3PO, Pepper (in the video below), which will be on sale in the US from February 2015 for less than $2,000.
Of course, they are very primitive, and there are much better examples of current humanoid robots – ASIMO, for example, on the right (will cost you millions). But for what is currently practical for us all to have in our homes – JIBO and Pepper are the perfect examples.
While we keep in mind that in the year 2000 most of us didn’t even have mobile phones, looking another 15 years into the future – what technologies can we envision? If Pepper is the Nokia 3210, then what is the iPhone 6?
The Dream Robot
I propose that it’s the dream – robots that can do many of the things we can; from mopping the floor, to washing the dishes in the sink. They will be battery powered, and on a full charge would run for at least a day (have a read about upcoming battery advances). They will be almost silent in their movement. You wouldn’t even need to give them jobs to do, because they will know what needs to be done, and when.
They will be trained in first aid… they will know more about health care than your doctor. Illness will be detected by them even before you know you are ill. They will help the old and disabled, and look out for the safety of our kids.
They will be experts at fixing anything; if they can’t physically fix something themselves, then they will guide you so you can do it yourself. They will even order the replacement parts automatically, being that they will be connected to the internet.
Controversially, they could be used as security guards, protecting your home from intruders. Continuing on this line of thinking, what’s to stop other robots being programmed to rob houses and banks? But that idea doesn’t belong in a dream, so for this article I am going to completely ignore it.
Planting flowers! They could do that! They could do all your gardening for you, with all of that waving around of sharp implements and revving chainsaws… to make all of your trees pretty and trimmed!
Okay, so maybe the nightmare version of this can’t be ignored. The idea of super-strength robots walking among us is pretty terrifying, at least according to Arnold Schwarzenegger, Keanu Reeves and Will Smith.
What needs to be priority is securing our robotic future as a safe one: When a robot takes your job, a new job will be created in robot safety, security and regulation.
If this dream is reached, the technology will be the most disruptive that humanity has ever created. It will revolutionise every aspect of our lives. Jobs, of course, is the first worry that comes to mind. Robots will save us money though (doing jobs we would have otherwise paid a person to do). And economies will benefit as our elderly will be cared for by robots: Europe recognises that it is expected to reach a 2:1 ratio of workers to pensioners by 2060.
Our economies will be forced to completely revolutionise. People won’t need to work as many days per week to survive. Researchers remain deeply divided though about whether robots and AI will make our economies bleaker or brighter; there is a great article on the subject here by The Atlantic, “The Robots Are Coming, but Are They Really Taking Our Jobs?“. And you can read more about the views of Google’s CEO on the future of jobs, here.
Robots and AI will revolutionise health care, with us all having our own personal doctors and carers in our homes. With robots able to take physical care of the elderly, we will be free to concentrate on making their lives not just comfortable, but enjoyable, by having time to organise social events and entertainment.
Many other industries will be effected. The better the robots become, the more trades they will be able to do. All of these jobs will be among the first to go – cleaners, window cleaners, factory workers, low skilled cooks (at fast food restaurants, etc), pot washers, shelf stackers, construction labourers, etc. As the robots gain in skill, more jobs will follow – Chefs, tilers, plumbers, electricians, builders, painters, carpet fitters, mechanics, delivery drivers, taxi drivers, bus drivers, accountants, etc. Jobs that will be protected are the ones where we will always want human interaction – in shops for example, in sales, customer services and waiters/waitresses. We will always want our nurses to be human, even if our surgeons and health advisors are better replaced by robots.
An overview about the future of robots and jobs can be seen in the video below, Humans Need Not Apply.
Policing and security will become massive growth industries, with robotics becoming a main area of concern. The weaponization of robots and AI will become an existential risk. We will need to have strict regulations regarding the use of robots. For example, do you want a stranger’s robot with you in the changing room at the swiming pool? Because they will be a walking video camera, and who knows who is watching. They will be capable of recording everything everywhere constantly. And that’s a tiny security concern relatively. Hacking will become a massive danger. Security needs to be able to ensure nobody can hack and control other peoples robots. Not only that, we need to be unable to order our robots to do illegal acts. And then there is the whole evil AI taking over the world thing – yeah. We need to embrace the security challenges to continue to innovate… while surviving. Read more about future security and surveillance challenges.
Reaching the Dream
Robots capable of doing so much, and at the same time being affordable for all – that’s a huge evolutionary leap from any of today’s robots. Huge technological advances are needed in 3 main areas – artificial intelligence, robotics and batteries.
We are beginning to see huge developments in robotics related technologies. Below, we explain the latest breakthroughs, and consider how long it will be until our dream robots will be in our homes.
Artificial Intelligence for Robots
We are not going to want our robot assistants to be conscious in the same way that we are, because that is when they stop being described as assistants, and start being described as slaves. Plus, self-aware robots aren’t going to let us use them as slaves.
Decisions that robots need to be able to make: What we need are robots that are capable of understanding the world, and able to use information to make their own decisions, rather than having all of their actions pre-programmed. They need to be able to react to a situations logically, to understand abstract concepts and make inferences when necessary. For example, if you tell your robot to “throw away that paint”, you mean for it to dispose of it properly and safely; the robot needs to logically understand what you are most likely telling it to do. If it does not understand the command properly, it may literally throw the paint, maybe all over your cat. It needs to understand your tone of voice and mood, to know if in fact you did tell it to throw it on the cat, so it can respond by calling you a lunatic. It needs to understand that it would not be the best idea to pour the paint in the bin, incase the bag leaks or splits. It need to know that paint needs to be disposed of in a special way. Logically, the best thing to do is to leave the paint tin with the bins, ready to be taken away by the local disposal services. Ideally, it would even ask you if it is a good idea to post on your social network, to ask if anybody else wants the paint.
Machine Learning – Current Efforts: The ability of machines to make decisions and learn, is described as machine learning. Computer science professor Andrew Ng defines it as “the science of getting computers to act without being explicitly programmed.” Currently, the company leading the field in machine learning is Google; mainly through its acquisition of other companies. Early 2014, Google spent $400 million on the purchasing of an artificial intelligence firm from the UK, called DeepMind. Find out more about Google’s efforts in machine learning. In October 2014, DeepMind announced that they had created the first “neural turing machine” – a computer that is a type of neutral network, able of mimicking the human brain’s short-term working memory. Basically, it is another step forward in the creation of an artificial human brain, read more here.
Google’s self-driving cars are an example of their current application of machine learning. They are not programming the cars to know that a red light means stop, they are instead creating a system that can learn that it means stop when it observes that when the light is red, humans always stop. So far their self-driving machine learning system has learnt from 700,000 miles of driving data. Eventually it is set to become the perfect driver, and massively reduce deaths on the roads.
The chart to the left shows an analysts prediction about the upcoming prevalence of self-driving cars. It is predicted that by 2026 it will be normal to see self-driving cars on our roads – you can read more, here.
Ray Kurzweil, a director of engineering in Google’s machine learning and language processing department, believes that AI will pass the turing test by 2029. To pass a turing test, an AI needs to fool a human into thinking it is another human, via text chat. The turing test is controversial though because of its simplicity, and some believe a better test exists, called the Winograd Schema Challenge. The challenge is a test of common sense; the AI will need to judge and learn from situations in the same way that we do, in order to pass the test.
There is a big leap between a robot understanding what it needs to do, and then being able to do it. And so far, they can do very very little, at huge huge cost. Every year robots from different companies are put through a competition called the DARPA Robotics challenge. Below shows the highlights of last years challenge. As you can see, there is a lot of work to do before they become anywhere near as agile as we are.
More impressive though is Asimo; in the video below, they showcase the latest version, as of June 2014. It is described as the world’s most advanced robot. The creators of Asimo do not put a time frame on when their robots will be in our homes. However, in an interview with a middle aged man, they told him that their aim is to have their robots in his parents home to help them when they are old. A little hint then, that they are optimistic their robots will be commonplace in around 20 years time.
Watching Asimo in action, we can see the dream on the horizon. The cost for Asimo though right now is $2 million. And for your money, you wouldn’t get very much at all; no way could it make you dinner; it couldn’t even bring you the mail.
Colin Angle (pictured left), CEO of one of the biggest robotics companies, iRobot, is disappointed with the rate of progress in the robotics industry. The reason, he says, that research and development companies are struggling, is because they are not focusing on developing products that have immediate consumer uses. He says that researchers are too often too attached to specific projects – they want to perfect a specific robotic ability, even though that ability would not have an immediate use for any consumer. The result of this, is a waste of money; millions spent on R&D, but no real short term profits. As result, this does not attract investors, and acts to slow down the rate of progress in the whole industry. Read more about Colin Angle’s opinions, here.
There are though success stories. Robotic products are being created and consumer uses are being found. The iRobot company are doing well in sales of their robotic hoover (for sale on the right). Of course, it’s not a humanoid robot, but it does provide them with a source of income so they can continue the R&D for new types of robots. Plus, the technology in it, especially the iAdapt sensing system, will contribute to the technological advances that will be of use in the robots of the future.
Google is able to ignore the economic rule of only developing robots that would provide an immediate consumer use. Over the past year Google has acquired many of the best robotics companies. All of these companies now don’t need to worry about finding quick consumer uses. Instead they can focus on perfecting their own long-term projects. For example, one of these acquisitions is Redwood Robotics, who specialise in making robotic hands and arms: They can now focus on perfecting the intricacies of the robotic hand, rather than worrying about rushing a product to market to avoid bankruptcy. We therefore have much less delay in research and development.
And with Google owning all the companies, it is in all of the companies interests to work together to create shared products, rather than competing and hiding technologies from one another. The downside though is that they operate in much more secrecy; they do not announce their individual breakthroughs publically, because they don’t need to do so to gain investments and hype. Google will want to keep breakthroughs secret, so to not give their competition the heads up about the direction the company is heading and on what timeline. All of which makes it hard for outsiders to estimate when products will be brought to the market.
Hints though have been given about the timeline. Edsinger, a robotics director at Google has said “my hope is that we’re going to see as much energy and effort pooled into robotics startups in the next 10 years as we’ve seen in social media in the last 10”.
Things are heating up; robotic tech is beginning to wow, and what has been covered in this article is only the tip of the iceberg. Exoskeletons are being created to get the paralyzed back on their feet, and give soldiers a much needed boost while carrying heavy equipment. A robotic arm called ‘Luke’ is being developed, to replace lost biological arms. Robotic gloves are being developed to assist those who are losing dexterity in their hands. Still only on the tip of the iceberg. Each year that goes by, we are seeing an increasing amount of breakthroughs. Next year we will create a sequel to this article, which will show only the robotic and AI advances that we will see during 2015. Follow us on Facebook to keep updated.
Upcoming Battery Advances
Batteries are long long overdue an upgrade. We want our robots to work for more than an hour before needing to go and charge for the rest of the day. There has been many announcements recently about breakthroughs in different battery technologies. There is no doubt that by 2030, batteries will be advanced far beyond their current measly selves. The latest announcement was for a battery which has a 10,000-cycle lifespan, and will charge to 70% in only two minutes. They will last the same amount of time on a single charge though, but still – usefull. They are set to be on the market in 2 years, more info here.
In April 2015, the ability to build microbatteries was announced. Using 3D-electrodes, they will shrink current batteries 10 times, while offering the same power. Therefore of course, the same size battery would be 10x more powerful. You will be able to jumpstart a car with your smartphone. More info here. And another plus is that the cost of our current Lithium batteries are falling exponentially. They are now half the price they were in 2010, and by 2020 they are estimated to be half the price again (chart to the right).
Solar Power Advances
Charging millions of humanoid robots requires our countries to produce much more power. To meet the power demand, we need huge upgrades in solar power, and a helping hand through a breakthrough in fusion power would be nice. Yup, that is exactly what is currently in development. Nuclear fusion energy may be available in as little as 10 years. More likely though are solar cell upgrades, which are coming fast. The cost of solar panels is dropping exponentially. Every year we are seeing an increase of 30% for energy which is supplied by solar power. Ray Kurzweil believes the US will meet 100% of its energy requirements from solar power within just 20 years. Even if we are anywhere near that number, our energy needs will easily be supplemented by other production means. Excellent article here about the upcoming solar power revolution.
Where is my robot already!?
So, hopes are up high and they are not coming down even when I’m senile – I’ll have a robot even if I’m paying my carer to walk like one. As early as 15 years is the optimistic time frame -the breakthroughs and trends described above certainly show we are set on target. What we don’t know though is what hurdles stand in the way. If you are thinking of any, comment below and start the conversation. One major hurdle will be the amount of regulation needed to allow for an army of superpowered gods to be allowed to walk our streets.