The potential for robots and artificial intelligence to replace humans and the effects of automation on jobs in general is raising many concerns.
The fear is that as a result of the advances in AI, production processes will become more automated. More intelligent machines and industrial robots that are capable of performing routine tasks faster and more efficiently than humans, could ultimately replace millions of workers in manufacturing, logistics and transport. Even in the service industry, many tasks could be made redundant by modern algorithms and smarter software programmes.
Research has shown that in the past decades, industries that have invested heavily in industrial robotics did indeed see the replacement of workers, especially in the US, with one robot on average substituting six human employees. The operational costs of robots are between one-fifth and one-eighth of the hourly cost of a worker in Europe. Even in China, robots have become cheaper than labourers.
Yet the process of machines replacing workers is not new. It has been ongoing since the first industrial revolution 200 years ago. A 2017 study in Germany showed that in the country that has the most robots per working population, robots had no effect on aggregate employment and that robot exposure increases the likelihood of workers staying with their original employer. However, this result was supported by strong worker representation in labour unions and fewer young people entering the manufacturing sector.
Robotics experts speaking at the Cayman Alternative Investment Summit earlier this month emphasised that it is not always a question of “robots or humans” but increasingly one of “robots and humans”. For the time being at least, robots are replacing humans mainly where the tasks are too dangerous or too mundane for workers.
Humans and machines
Australian robotics and artificial intelligence pioneer Robert Brooks said the advances in deep learning and machine learning driven by the rapid increase in computer power and Moore’s Law (processing power doubles every two years) have been remarkable, but they are often over-generalised and not the answer to everything.
The former Massachusetts Institute of Technology robotics professor said, “I think we are still at the very beginning of the journey.”
Brooks expects hybrid systems of robots assisting humans to become more commonplace.
“Humans are still going to be very much in the loop for quite a while to come,” he said.
Just three or four years ago, the assumption was that radiologists would ultimately be replaced by AI software programmes. Today, the software, which can quickly highlight anomalies, is seen as another source of data for radiologists, who are able to handle more cases, more quickly, Brooks noted.
In fact, when humans and machines analyse radiology images together, research has shown that the accuracy is higher than for either radiologists or AI programmes working on their own.
He argued that “humans and machines” should be the way we think about it, for now. “If we go out 50 years or 100 years, who knows, but for our purposes we should look at humans and machines together.”
As robots will increasingly enter the lives of people to assist or share tasks, it will be important that they can be easily re-tasked by users or workers, Brooks said. “It cannot be some IT department that tells them what to do. It has got to be people interacting with them.”
Currently, robots are often deployed to carry out low-interaction, highly repetitive and monotonous tasks that do not appeal to humans. When humans and robots share these types of tasks, there is a real risk that workers can become overwhelmed.
Steve Toebes, chief robotics officer at HDS Global, a grocery and general merchandise delivery service, said systems need to be designed in the right way. If robots deliver items to a human picker, like at Amazon, one of the biggest users of robots, it will be difficult for the workers to keep up with the pace set by the machines.
“You can get burnout, if you have the wrong system,” he said. But with the right set-up, humans and robots can work together.
In the logistics field, robots are also needed to cut costs. While online food ordering is growing on a massive scale, companies are struggling to make money, Toebes said. To offer in-store pricing without delivery fees, online food retailers will have to use robotics and AI to control costs.
At HDS, that means creating fully automated warehouses that combine large-scale automation with robotics. Traditional conveyors are replaced with more flexible and cheaper mobile robots. These individual robots work together using data analytics and artificial intelligence to determine the best way to run a warehouse.
The company is also replacing human pickers, mainly because it is difficult to find enough workers.
Yaro Tenzer, co-founder of RightHand Robotics, agrees that warehouses are struggling to meet the demand for online retail orders amid a labour shortage.
“At the moment, the way they are deploying labour is really challenging,” he added. “When the robot is fully functional and operating at a fast rate, you really don’t want to be next to that machine.”
Where the collaboration can come in, Tenzer argued, is when humans step in either remotely or physically to help the robot resolve a problem.
While robots are still largely used for dirty, dangerous or dull jobs, the so-called three Ds of robotics, that will change as artificial intelligence improves. New potential-use cases for robots are appearing everywhere.
One area where robots could collaborate intimately with humans is care for the elderly.
As changing demographics rapidly reduce the number of working people for every retired person, finding enough care workers will become a growing problem. In some countries, cheap immigration can be a temporary solution to the labour shortage in elder care, but Brooks believes robotic assistance will be needed in the long term.
“I think we are going to see robotics and elderly people working very closely together.”
This will not necessarily require complex AI or “robot friends”, Brooks said. “For instance, my mother, the limiting factor for her going into elder care was that she could no longer get in and out of bed by herself. If she had been able to get some assistance to getting in and out of bed, she could have lived independently a lot longer.”