With fears of automation-induced mass employment growing we need to act now to shape a robotic future that works for us and not against us write Prof Noel Sharkey and Dr Aimee van Wynsberghe
Barb the builder whistled on her way to work today. Her new exoskeleton suit has arrived and her back will be saved from lifting and laying bricks to build houses. But the happy mood took a nose-dive when she read about Hadrian X in the newspaper. It’s a new bricklaying robot that can lay 1,000 bricks an hour. It could build a house in 2 days that would take Barb and her team 6 weeks. ‘How can we compete with that’, she thought.
And Barb is not the only one who should be worried about robots taking her job. There has been a rapid upsurge in the number of robotic applications planned to replace human workers. Robots have been operating in factories since the 1950s but now they have moved outside into the service industry to serve food and drinks, to make burgers and pizzas faster than any human, to clean up and to take over transport with autonomous cars, trucks and busses. The mining industry is being transformed with robotics as are agriculture and delivery services. And this is just the beginning.
Reports abound about AI and robotics creating mass unemployment within a couple of decades. One of the earliest studies by Frey and Osborne (2013), forecast that 47% of all jobs in the USA could be computerised within twenty years. Predictions were based on criteria including repetitiveness and social interaction. Deloitte (2014) followed with studies in Britain, Switzerland and the Netherlands showing similar results. And more recently the Banks of America, UK and Italy have expressed grave concerns.
Is there a solution? No, but there are a number of suggestions. The Dutch report Mastering the Robot. The Future of Work in the Second Machine Age suggests an inclusive robotisation, with the core concepts of complementarity between human and robot skills and capabilities and, ownership of work. McKinsey & Company, suggest that we should rid ourselves of the concept of ‘occupations’ and instead utilise the concept of ‘work activities’. Robots may be used to replace or assist specified portions of a job rather than replacing or assisting an overall trade or occupation. They argue for a five-factor approach to predicting the automation of particular activities. Among these are the benefits of automation and the cost of workers who might otherwise perform the activity. This leads to a conclusion of only 5% displacement of workers. However, it does not take into account the plummeting costs of robotics and the accelerating number of tasks that can be performed.
Bill Gates has suggested that companies pay tax for a robot replacing a human. This would then be used in some way to compensate redundant workers. While tax sounds like a good idea it is not an ideal solution and is fraught with difficulties such as defining a robot and how many jobs it actually displaces. For example, it may be a new automated robot system that never had human workers like the Ocado supermarket that uses a swarm of co-operating robots for picking, packing and sorting groceries. And there are serious concerns that a robot tax could slow down this sort of innovation and possibly create more unemployment over the short term.
Providing a universal basic income (UBI) is another widely heralded suggestion. This takes a number of forms such as giving everyone the same amount and then taking some away for every dollar earned. This has difficulties as well. Who will set the basic amount and can it be guaranteed when new governments come into power? And where will the money come from? With significantly fewer people being employed we will need new economic models. We need to think about different possible scenarios and the likely impact they could have on our basic human rights and dignity. What kinds of inequality could emerge with a new technocratic elite? Will it aggravate gender, age and ethnic biases?
We need to decide what kinds of responsibilities we as a society will allow companies to delegate to a robot or AI
A recent head-in-the-sand vote at the European Parliament went against both robot tax and UBI without offering any alternative or promise to investigate the issues further. Other countries are less forward in their deliberations. Yet this is an urgent matter as the potential for a robot revolution in all areas of the workplace unfolds over the next decade.
Whether or not mass unemployment is immediately on the cards, we need to decide what kinds of responsibilities we as a society will allow companies to delegate to a robot or AI. We currently have no policies in place. Society needs to be better informed and play a larger role in decisions about the kind of good life we want to have and how robots fit into this picture. We need to shape a future with robots worth wanting.
By Prof Noel Sharkey (co-founder and chair of the Foundation of Responsible Robotics – FRR) and Dr Aimee van Wynsberghe (co-founder and President of FRR)
I’ve studied robots and automation 30 years ago. And those questions were asked for ages. Since I left university I haven’t been in touch with any roboter at work. I’ve seen an autonomous grass mower. But that’s about it. Even the Christmas parcel sorters at Royal Mail are human and get paid minimum wage, if they are paid.
I remember the same sort of discussion back in the 70s when computers were beginning to come on line, the argument went that computers were going to take over and we would all end up unemployed or in menial jobs that computers could not do. The other part odf the argument was that by using computers we would enter a new age of prosperity where we would be working minimal hours and would have more leisure time than we could dream of. I am still awaiting the apocalypse.
Like most modern things they will have a place in industry and the wider world but we must be careful that we do not become modern day luddites and condemn technology purely because it is something new or different to what we are used to.
Whilst I absolutely agree about avoiding being Luddites, I think we need to be very careful to be reasonably even handed in distributing the fruits of these robotic enterprises, to avoid a detached “under class”, who see no benefits, and reject them (possibly violently) – in recent time, Brexit is a prime example of where the “masses” felt that the EU (& globalisation in general), did not benefit them, so voted to “opt out”
The robots will be too late, the politicians wil have stolen all the rights, jobs, and dignity, along with all the money long before a robot can get there!
I think Robots will become part of the working world that is very similar to the personal computer.
Come in Robots !!! Major problems in today’s society include an aging population and high expectation from all workers. Robots will hopefully reduce the crisis we might otherwise have from a shortage of workers as people retire. Workers on minimum wages will be replaced by robots releasing them for higher paid tasks.
I look forward to a robot carer in my care home instead of an underpaid foreign immigrant who I feel guilty to have brought to this country only to be paid barely a living wage.
With robots helping productivity we could have shorter working hours and so more time for soccer and golf ! What’s to be afraid of ?
Given that there are many millions of synapses in the human brain, and also that science is just begining to understand that each synapse is a form of CPU, it will be a long long time before we have anything even approaching artifical intelligence. While it may be possible for humans to design robots to build robots, humans will have to design, and program them, and also maintain them. Humans are very versatile, whereas robots are very good at doing repetive tasks quickly and accurately. The house building robots are a bit of a red herring, often designed and built, but never actually used, because of course they are phenomenally expensive, and among the turmoil and dampness on an exposed building site, totally unreliable, and of course, very expensive to repair. Watched a “3D house printer” spraying concrete to build houses some time ago (in a lab, not on site) which seemed to be ignoring the fact that concrete is the most awfull material to build houses from, cold, damp, doesnt last! Witness the number of concrete houses and blocks of flats being demolished. So a robot could build a house on two days? No it couldnt! Six weeks is the time needed to complete a house, this robot only lays bricks, it cannot fit windows and doors, or roofs, install plumbing and wiring, or even plaster walls. there ARE robots that can plaster and render walls, but can you really imagine getting one to the upstairs of a house? Build a house in 2 days? Where does all the hundreds of gallons of water in the mortar and plaster go! Houses need time to dry out as they are built. It simply Won’t work! The point arrives where a human being is not only more versatile, but also cheaper too.
Robots are here, any car production line will demonstrate how they can be integrated into production. There are still manual operators as some tasks still require a degree of dexterity and vision that “can be automated”, but at a cost..
Cost and materials technology will drive the whole programme.
1000 bricks an hour is great bricklaying, but only if you are building a long wall. The cement has to go “off” before it is too heavily loaded, but I’m sure fast setting cement could be used for a fast laying robot!!
The “solutions” for providing the lowest cost of goods and services have been around since the nineteenth century. It should be no surprise, that when the total cost of human activity per unit produced exceeds that of mechanized or robotic solutions, the higher cost option will be supplanted by the lower. This is not a cause for alarm, but should be celebrated as the wealth in society continues to increase as a result of higher productivity and the lower costs. If the “social” objective is to provide jobs for humans then the obvious solutions would be to: 1) address the root causes of the high cost of labor, and 2) adapt to the changing circumstances by improving the skills of human workers. One could accomplish the former in large part by abolishing minimum wage and licensing laws and the latter by training workers in the requisite knowledge that would support automation, robotics, and computer programming.
We don’t seem to have a problem when computers and smartphones taking our jobs, privacy, social interaction and national security, so I’m sure this won’t be a big deal either.
If we don’t automate then we’ll never be living in an equal society. I have healthcare, communications, electric lights, entertainment, libraries of knowledge at my fingertips that no rich person could have had for any amount of money 100 years ago. The difference between me and one of the elite is enormous in terms of money but not so much in terms of lifestyle, I think. Someone or some thing has to provide all this stuff cheaply enough for me to be able to enjoy it and at the moment that’s the poor people in other countries plus some degree of automation at home. What happens when there are less and less poor people to “exploit” as they get wealthier? We will need more machines then.
I was struck by the way the title is posed: “Will Robots take our rights, our jobs and our dignity?” I don’t think robots are capable of taking anything. Rather, for one reason or another, somebody will give them these things. We need to think about those people and those reasons…
Strange how it always seems to be the manual jobs? Surely a robotic account, lawyer or politician would be easier to build?
The fact these questions have been asked for ages, or that robots also seem to have a lot of promise of bettering the world, don’t take away the necessity to think the design and implementation through.
Possibilities such as shorter workweeks or better welfare could be great results for the increased use of robots, but in that case we do need to make sure policy reflects that. Just building and implementing robots will not solve these issues, making sure they are build and implemented correctly just might. That is where I see the benefit of thinking about these issues in robotics..
Science fiction often becomes science fact – just look at the writings of Jules Verne, for example. No doubt robots could do some things very well, but their widespread use will cause resentment and lead to civil unrest. Has nobody seen the many ‘robots taking over the world’ films? The more intelligent something becomes, the more dangerous it is. Man is a prime example. Cro Magnon man could not have conceived of such devastating power as the hydrogen bomb, yet more than once we’ve been on the brink of Armageddon. You create Artificial Intelligence and that evolves into Real Intelligence. Those intelligent, self-replicating robots wouldn’t need humans and they would eradicate us. We’re bound to wipe out life on this planet somehow or other; perhaps robot overlords will be the method
If you’re looking at the dystopias, you might look at the sci-fi where AI and robotics are beneficial as well: Iain M Banks et al are good examples. As our resident sci-fi author Jon Wallace says, writers like dystopias because conflict is interesting to write and read about, not necessarily because they’re most likely to happen.
Sci Fi all seems to look at the human / robot side of things (‘cos how many robots are likely to read sci fi for leisure?).
However, if you were a non carbon based, non air breathing entity on this planet, looking for the resources to exist and expand and evolve, the first logical step is to get off the planet. Minerals and other resources aplenty already in orbit, or available on the moon. Solar energy much more efficient away from atmosphere and planetary rotation. No need for food, or life support, travel time becomes largely irrelevant.
Colin.
Aids to working more efficiently have been around since the first man looked at the local stream and thought he could use it to make his job easier. What has changed is the pace at which these applications are developing and causing us to think more about the consequences and solutions.
Dignity isn’t taken by robots but politicians and lawyers.
Will robots buy that car?
For robots to be profitable, someone needs to buy the goods and services.
If people lose their income due to unemployment, will they need to generate that profit by taking out loans?
Would it be easier if the government just prints the money and gives it to corporations without involving workers at all?
Thank you for all of your great comments. There seems to be an emerging consensus here that there is nothing to worry about. If you base it on the 1980’s worries bout the PC and the rise of computing, you would appear to be right. I was one of the people in the 1980 saying that computing would provide more new jobs.
I also felt that the same was true about the upsurge in robotics – surely it will just move the jobs around. Then I was on the fence. But over the last two years, traveling around the world and seeing all of the developments (far too many to mention here) not just in robotics but in narrow AI, I fell off the fence. I realise that there will be new jobs created but there will probably be considerably fewer. Just because automation did not take the all of jobs in the past, it doesn’t mean that it will not take all of the jobs in the future.
But I have to say genuinely, that I really hope that I am wrong about the negative outcome. To a large extent it is in our hands as consumers but it is well to think about potential implications of new technology in advance and make sure that people are aware of the potential dangers.