It is a commonly held fear that robots and artificial intelligence will one day end the world. Or that, one day, these technologies will become so capable that we, their creators, will lose all control over them. With a strong level of confidence, it is concluded that a robot apocalypse is far off in the future, assuming it is even possible.
To set the stage, what is automation? What is this all-powerful force that seemingly threatens the jobs & livelihoods the world over? To be honest, it is less threatening once one understands the topic a bit better. Automation is the use of machinery to help with manufacturing and processing tasks. The word “help” is the operative word here. In the projected future, most of the integration of machines into the workplace is to help with tasks that humans are performing already. The future of automation looks more like cooperation between humans and machines, instead of a widespread machine uprising over the human populous.
People have feared the rise of automation since it was first popularized. There are roots in this fear that go back to the the 1920s with Karel Čapek’s play R.U.R. ‘Rossum’s Universal Robots’ (1920)[1], which coined the term “robot”. It comes from the Czech word “robota” which means “forced worker”. This highlights the ingrained fear of the rise of technology and the potential to replace humans in the labor force, since its beginnings in cultures the world over. More recent hallmarks in pop culture have only exacerbated this. In Terminator (1984), an artificial intelligence called Skynet literally plays the main antagonist for the entire franchise, using its robot forces to threaten the end of humanity. The story of 1987’s Robocop revolves around a police officer, who is cybernetically enhanced himself, fighting a robot meant to replace police forces in Detroit, Michigan. The droids in the original Star Wars Franchise come to literally take over the military and labor for the galaxy. These are known as some of the most famous films of the past century and all of them perpetuate the same fear of technology and the usurpation of roles in human society as we know them.
However prevalent in modern culture the idea of robots taking over may be, one must remember that everything from Capek’s R.U.R. ‘Rossum’s Universal Robots’ to any sci-fi show that has maintained the fear or automation has been fiction. As for the news reports and articles one can find online which exist in real life, it is found that they should not be taken seriously at face value. This fear will probably remain with many & will not go away easily. Going forward, it will be important to remember that machinery will be coming into workplace slowly. We are far from the reality of people walking into work one day to find machines in their places & they have no job left to perform.
When most new reports have gone on about how many jobs would be taken up by machines, these claims have been debunked, unfortunately after they had already been released. These claims were actually representing the that machines would perform some of the number of tasks one has as part of their job. Just like how a teacher has other tasks to their job besides teaching students, such as grading papers and attending meetings, so do many people who work in the industries that face the integration of machinery. For example, one on the assembly line may have a task of inserting screws into a car, on top of the rest of the tasks to assemble the car itself. Instead a machine would handle the insertion of screws, leaving the person to be able to dedicate more time and effort into their other tasks. So machines may be added to manufacturing processes and the like, but this does not necessarily mean that entire jobs will be replaced. This miscommunication has been cited and attemptedly clarified by many institutions. Boston University published a paper titled “Automation and Jobs: When Technology Boosts Employment”[7] which stipulates this clearly. It adds that 36% of tasks in the occupations they studied are “potentially automatable”. There is a potential for these tasks to be supplemented by a machine. The paper also shows that people’s tasks would then shift to handling more quality control and other segments that ensure better performance from the assembly line instead of faster output. So instead of automation causing mass unemployment, it is much more likely that we would see people focusing more time on a smaller pool of tasks.
On top of this, there are machines that will actually enable more jobs for people. With the argument that humans will be replaced by machines in the workplace to do their jobs, one should also think about the jobs to create & maintain those machines. The Mckinsey Institute[8] published an article stating that because of technological advancement, we can expect 8-9% job growth worldwide by 2030. For example, in this day and age, we are not sophisticated enough to have machines that can efficiently reproduce other machines. This means there will be people needed to design and produce them. In addition, automation may make tasks easier to perform and with greater precision, they still require humans to operate these machines. This means that even with the jobs that do get lost from the integration of automation, there will more jobs produced to mitigate that loss.
One of the major advents thrown around, when the topic of automation is discussed, is that artificial intelligence will be the “be all, end all” of humans working as we know it. But what isn’t usually brought up is that humans will be still be needed to produce something like an artificial intelligence. Ben Medlock PhD, and founder of multiple successful tech startups, states that artificial general intelligence (AGI), the level of AI (artificial intelligence) & machinery that the human population is worried about, is “decades behind, and therefore decades away”, according to a Forbes Article[9] on the possibility of AGI in the future from August 2018. The progress made into AI that we have heard of so far has been on machines made to serve 1 purpose. Google’s AlphaGo was designed to only play the game Go. IBM’s Deep Mind was made to only play Chess. Elon Musk’s OpenAI project only plays DOTA 2, a massively popular online video game. These projects are only the most famous of the AI projects that have been undertaken in the world. All of them, have been painstakingly developed for 1 specific purpose. We are far from a AGI, that can do more than just play 1 game or serve a single task. The general population has little for worry about for the near future.
Moving from technology that has or will be integrated into the workplace, a business has to choose to employ automated machinery in the first place. First, there is an enormous upfront capital cost to automated machinery. One has to pay to develop one that can perform a desired task, including the material, production, and training costs to get the machine up and running. According to the Encyclopedia Britannica[10], such costs for an automated machine are immediately measured in the millions of dollars already. Not to mention that if something were to go wrong in the work environment, like a mishap on the assembly line, a machine is cannot be trained to handle any such eventuality. According to a paper titled the Ironies of Automation[2] from the University College of London, “[the] automation of industrial processes may expand… problems with the human operator… and on the potential for continued use of the human operator for on-line decision-making within human-computer collaboration.” Since these machines are trained to do one or even a few tasks at most, they cannot be taught how to react to changing factors in their work environments like humans can. Take the case of self-driving cars for example. They can only drive cars in the conditions used to train the machines. This is why we have seen a number of reports about such cars being destroyed by threats they cannot be normally trained for. For instance, a TESLA self-driving car collided with a commercial shipping truck at an intersection[12] despite even common sense stating that one should not drive through an intersection while a truck is moving through. Every human knows that the world is random & is full of unexpected happenings. A machine is very unsuited for many situations outside of a clean laboratory. And since mankind if decades away, at least, from producing a truly capable AGI, we are stuck with having to train machines for conditions that the humans that make them can predict.
There is something to be said about the long-term benefits of automation. A machine may be able to work 24 hrs a day, 7 days a week, & 365 day a year, without any need for sleep or pay. But to take full advantage of these benefits, a company would have to pay for their entire assembly line or work force to be automated. This is an enormous cost, both money and time wise. And if they were to automate to any extent, it takes years for the projected benefit to the implementation of automation to outweigh the cost. And like humans, machines have a maintenance cost. They have parts that need to be replaced & occasionally need repairs, just like how humans need to eat, and go to the doctor when they are not well.
With costs coming into the discussion, the economic component of automation comes more into focus. Take the fearful idea of the a future where machines have taken over the supply chains and manufacturing jobs. In this future, people still have to receive a wage to afford these now machine-made products. Given the mostly capitalist markets employed by much of the world, people will need to earn a wage by working jobs that contribute to their respective economies. Goldman Sachs, one of the most influential investment banks in the world, still holds consumer spending[11] to be the backbone of global markets. And people need wages & salaries to spend on the products & services companies release. Regardless of how aggressive the push for automation will get, we can rest assured that people will have to stay employed to continue to consume & keep their economies going.
The wave of automation is not as much of a fearful force as many believe it to be. The prospects for machinery and artificial intelligence replacing humans are low and much farther ahead of us. It is concluded that even with the integration of automation in the workplace, people will most likely see a shift in the tasks they focus on through the day, instead of their jobs being replaced. And in the instances they are replaced, their companies would incur a massive cost to do so & there are jobs that will come out of the rise of machinery that are open to the public. People have to stay employed to remain consumers of the products companies want to sell. Otherwise, they would have no customers to satisfy or keep up with.
Sources:
- Čapek Karel. R.U.R. Rossum s Universal Robots: Kolektivní Drama o Vstupni Komedii a 3 dějstvích: 2. Vydani. Penguin Random House, 1921.
- Ironies of Automation, a paper published by the University College of London https://www.sciencedirect.com/science/article/pii/B9780080293486500269
- https://journals.sagepub.com/doi/abs/10.1518/001872097778543886?casa_token=0I8U0kKi8ygAAAAA%3A5-J7RRgrdMJHcohl3DM5EqLKSi2Kb_6e6eYB9T6trF3N0xSdX8kz7BJsVGoGWEG0n-2-GIvhWl_oCw&
- https://medium.com/@autodesk/the-paradox-of-worker-shortages-and-automation-fears-5951b82fe100
- https://www.law.upenn.edu/live/files/3397-di-nucci-e-santoni-de-sio-f-whos-afraid-of-robots
- https://www.ft.com/content/a9ec6360-cf80-11e5-92a1-c5e23ef99c77
- https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2690435
- https://www.mckinsey.com/featured-insights/future-of-work/jobs-lost-jobs-gained-what-the-future-of-work-will-mean-for-jobs-skills-and-wages
- https://www.forbes.com/sites/forbestechcouncil/2018/08/28/how-far-are-we-from-truly-human-like-ai/#6cda424231ac
- Encyclopedia Article on Automation https://www.britannica.com/technology/automation/Advantages-and-disadvantages-of-automation
- https://www.cnbc.com/2018/12/09/goldman-sachs-consumer-spending-to-keep-hopes-on-economy-alive.html
https://arstechnica.com/cars/2019/03/feds-investigating-deadly-friday-tesla-crash-in-florida/