Will smart machines be able to do all the work that humans do? Even though, humans and machines have very different skills and strengths, there’s a good chance that machines will.
T.J. van der Schaaff, Quadruple Master’s Degree academic at Vrije Universiteit, Universiteit van Amsterdam, and Tilburg University, the Netherlands.
What is artificial intelligence (AI)?
Artificial Intelligence (AI) makes it possible for machines to learn from their experiences, adapt to new inputs, and perform human-like tasks. There is a lot of emphasis on deep learning and natural language processing in the AI examples you hear about today. The application of AI includes everything from computers that play chess to self-driving cars, for example. In practice, AI technologies can help computers learn how to do certain things by analysing a lot of data and looking for patterns in the data.
Why is artificial intelligence important?
In order to understand the societally relevant aspects of AI that are discussed in the paragraphs, that follow, it is rather essential to understand the underlying reasons of AI’s importance to humans and society in general.
The first reason for AI being important is, because it can perform frequent, high-volume, computerized tasks all the time, instead of automating manual tasks like humans would do. Second, it does so consistently, without having breaks, and without ever getting tired.
The second reason for AI to be of importance to humans and society is, because AI makes things that already exist, smarter. Many products you are using right now, are continuously getting better, because of AI technologies. One example is Apple’s voice controlled assistant called Siri, which was first presented to the public in 2011. Since then, Siri has developed the ability to handle follow-up questions and support language translation. Siri is also now an active assistant for macOS as well as watchOS, besides its original home on iPhones and iPads.
The third reason for AI to be of importance to humans and society is, because AI learns through progressive algorithms that let the data do the programming, so it doesn’t have to write the code itself. When AI looks for patterns and structures in data, it helps algorithms learn how to do things. As with chess, an algorithm can teach itself how to play. It can also teach itself what product to recommend next on the internet. When new data is added to the models, they change how they work.
The fourth reason for AI to be of importance to humans and society is, because, AI uses neural networks that have a lot of hidden layers to look at more and more data. In the past, it was impossible to build a fraud detection system that had five layers that were hidden from view. All of that has changed thanks to powerful computers and a lot of information that can be stored. When you train deep learning models, you need a lot of data because they learn right from the data.
The fifth reason for AI to be of importance to humans and society is, because the deep neural networks that are used by AI can be used to achieve great accuracy. Deep learning technologies are present in everything you do with intelligent assistants, such as Alexa and Google. As you use these products, they get better and better at what they do. When it comes to AI in the medical field, techniques from deep learning and object recognition can now be used to find cancer on medical images with better accuracy.
The sixth reason for AI to be of importance to humans and society is, because AI makes the most of the information that is hidden in the raw data that is provided to the mechanism. When algorithms can learn on their own, the data itself certainly becomes a valuable asset to possess for institutions and companies. It’s all in the data. AI can help you find them. Because the role of data is now more important than ever, it can give a company a big advantage in the market. If you have the best data in a competitive field, even if everyone else is doing the same thing, the best data will win.
Artificial Intelligence (AI) can be used in a wide range of industries, but can it be used in all of them?
AI is disrupting multiple fields of studies and work with its high level of effectiveness and quantitative abilities. Yet, there are many other industries, in which AI is not exactly that effective. In order to obtain the necessary insights in the relative susceptibility of various jobs across industries to increased levels of AI-based computerization, I will first provide information on the existing feelings of fear and contempt in people around the world regarding AI taking over all possible “human” jobs, after which I will discuss the possibility of AI-based technologies taking over “human” jobs across various industries.
Fear and contempt of AI taking over all jobs
Carl Frey, an economist, and Michael Osborne, a machine-learning researcher, both of the University of Oxford in the United Kingdom, warned in 2013 that 47% of US employment were at high danger of automation within two decades. According to others, the assertion was disastrous. “It was akin to declaring that all of the world’s volcanoes would erupt next year and that we will all perish,” Glenda Quintini, an economist at the Organization for Economic Co-operation and Development (OECD) in Paris, explains. The backlash spread to the artificial-intelligence field, where several researchers expressed alarm about their work’s possible implications.
There was relief when a 2016 assessment suggested a level of merely 9%. Quintini and a colleague released their own estimate this year, estimating that 14% of employment in the OECD are at danger of being eliminated. However, even these lower figures reflect a massive number of displaced employees across various industries, because of increased adaptation of AI-based technologies.
Yet, numerous economists argue that the issue is how to balance them. One counterbalance is that as automation gains traction, occupations will evolve to become more difficult to automate. Quintini discovered that employment in Norway was facing a rough one-third lower risk of automation than those in Lithuania, which was the result of the increasingly advanced levels of automation in Norway, and implied that most jobs had already been redesigned to include a higher social component. The inclusion of higher social components across various jobs in numerous industries was finally able to deal with the rising numbers of jobs that were suddenly filled by advanced AI-based technologies.
In the United States, such shifts in the abilities necessary for certain vocations are also evident. Between 1980 and 2015, David Neumark, an economist at the University of California, Irvine, discovered that the percentage of readily automatable labour done by low-skilled US employees decreased by 20%.
In which particular industries can Artificial Intelligence (Al) take over “human” jobs?
As, artificial intelligence (AI) is very quickly getting better at many “human” jobs, like diagnosing disease, translating languages, and giving customer service, reasonable fears of AI eventually taking over all human jobs in the economy, are on the rise. Yet, the fear of AI taking over all human jobs in the economy, is only one of the possible outcomes, and is not even that likely to be realized in the future, claims a trio of technology experts, consisting of Erik Brynjolfsson, Daniel Rock, both from Massachusetts Institute of Technology, and Tom Mitchell from Carnegie Mellon University, in their published research article on the economic consequences of artificial intelligence and robotics. In their research paper, the researchers point out that the impact of machine learning (ML), the self-programming, self-adjusting core of AI technologies will affect very different parts of the workforce than earlier waves of automation that aimed to replace human jobs. Instead, the present form of automation through ML will rather occur on a task-by-task basis.
The first example of an industry that may be taken over by the advanced developments in the field of AI are the taxi industry, and chauffeur services industry. Shelia Cotten, a sociologist at Michigan State University in East Lansing, conducted a study on trucking, taxi driving, and chauffeuring automation, in which she estimated that employment losses among professional truck drivers would be rather light, at least until 2030, with numbers in the low hundreds of thousands in comparison with taxi drivers and chauffeurs who were estimated to be most impacted in comparison before 2030.
The reason that the taxi industry, and chauffeur services industry have been estimated to be largely impacted in the near future, is because of the advanced development of self-driving automobiles by both technology companies, and car manufacturers, at a breakneck pace. Waymo’s test vehicles have logged more than 10 million miles on public roads, and the firm launched its commercial operations of a fleet of autonomous taxis in Phoenix, Arizona back in 2018. In addition, there’s General Motors-backed Cruise that is currently seeking its final approval to commercialize robotaxis in San Francisco, and Ford motors, that has quite recently launched their first commercial self-driving service, successfully, and is currently focusing on what it will take for customers to embrace the self-driving car technology in order to build trust and keep coming back in the future.
In 2018, Waymo LLC, an American autonomous driving technology development company, and a subsidiary of Alphabet Inc, which is the parent company of Google, introduced their self-driving taxis in Phoenix, Arizona.
The second example of an industry that may be taken over by the advanced developments in the field of AI is the manufacturing industry, and its sub-industries. The Institute for Supply Management Manufacturing (ISM) has been using their monthly business reports to discuss the accelerated growth rates in numbers of fully automated production facilities in the manufacturing industry, because of the rapid adaptation to AI-based technologies that are cheaper, more capable, and more flexible than human workers in manufacturing industries.
The invention of robots has revolutionized industrialization and production processes in all sectors of the manufacturing industry. Though manufacturing was considered to be the most mechanized sector of the manufacturing industry, further application of completely autonomous AI-defined robots in automotive production facilities remained a long way off for manufacturers. However, right now, AI-defined robotic solutions are finally transforming all of the production processes in these production facilities. These AI-defined robotic mechanisms are characterized by increased dexterity, and autonomous learning capabilities that could potentially transform entire manufacturing processes and the broader landscape of the manufacturing industry, and its industry sectors.
These rapid developments in innovative AI-defined robotic technologies are gradually reducing the demand for “human” labour in all sectors of the manufacturing industry. The automotive industry, for instance, has experienced a tremendous rise in the use of AI-based technologically advanced robotic technologies. This tremendous rise will, inevitably, affect the total number of employed “human” workers in the automotive sector of the manufacturing industry.
To give you, the reader of this report, an idea of the large-sized scale on which these new AI-based technologies may be produced for application in the automotive sector in the manufacturing industry, I will provide you insights in the current demand from the market for robots that can be used in the automotive sector of the manufacturing industry. In recent studies on existing automotive assembly robots, a group of market researchers from Boston Consulting Group found that a total of 35% of the total number of robots, that are manufactured all around the world, are made for application in the automotive industry. In particular, the highly industrialized countries, which have major automobile manufacturers and automobile production plants, were found to be the final consumers of the current robotic technologies for manufacturing motorized vehicles. This might imply for the near-future that, from the moment when completely autonomously operating AI-based robotic technologies will be applied on a large scale in automobile production facilities around the world, “human” jobs will be lost, just as before the initial introduction of robots in manufacturing plants.
The increased use of robot arms in car manufacturing led to job losses. Credit: Luke Sharrett
However, there is no assurance that people’s occupations will follow suit. Instead of worrying about potential job losses, executives should help cut down on jobs where AI and ML do the boring work while human workers perform higher-level tasks.
What is the overlap in tasks and work between the industries discussed above?
In theory, artificial intelligence (AI) has the potential to impact practically every industry and professional category. Yet, a recent research article that was published by Brookings Metro on job automation processes through adaptations of AI-based robotic technologies, indicates that less-educated, lower-wage workers are most vulnerable to displacement, because of the new AI technologies. In particular, employees working in the manufacturing industry are at risk of losing their jobs, as they become highly trained in a variety of jobs and are already closely connected with AI-based technologies that are applied on a large scale in manufacturing facilities.
However, the authors of the Brookings Metro article expect advanced AI-technologies to have the greatest impact on men, prime-age employees, and white and Asian American workers. Larger, more technologically advanced metropolitan regions as well as villages with a strong industrial base will most likely experience the greatest disruptions from AI-based technologies.
A visual representation of potential takeovers of former “human” jobs by AI-based technologies in the workplace.
Which jobs cannot easily be replaced by AI?
Yet, luckily for people, there are many other industries where Al is not exactly that effective, as in the industries that were mentioned in the previous paragraphs. Be it a Human Resource manager of an organization, a Public Relations manager or an artist, there are industries, which are characterized by task-activities that require a greater amount of human sympathy, experience, and interpersonal skills, which are significantly harder for AI-based technologies to perform compared to performing routine, quantitative, mathematical tasks.
The emergence of artificial intelligence, in particular, would fundamentally alter labour markets, a reality that many employees are concerned about. There are four types of employment, however, that will remain unaffected by the artificial intelligence revolution, according to Dr. Kai-Fu Lee, who is the Chairman and CEO of Sinovation Ventures, and President of Sinovation Venture’s Artificial Intelligence Institute. Lee said in an op-ed for Time headlined “Artificial Intelligence Is Powerful, and Misunderstood.” “How Can We Protect Workers?” was published. They are as follows:
First, the creative category that includes jobs like scientist, novelist and artist. These creative jobs are hard for AI to perform, because AI needs to be given a goal to optimize, instead of inventing a goal itself, states Dr. Lee. While this is true, artificial intelligence employed such optimization in 2018 to construct a picture of a fake person. Obvious uses neural networks to scan hundreds of photographs and then created a new image using the information. The resulting work, “Edmond de Belamy, from La Famille de Belamy,” was sold online by Christie’s for $432,500.
Second, there are positions that are complicated and strategic in nature, such as executive, diplomat, and economist. The intricate requirements of these types of work “far exceed” what computers are capable of processing, he adds.
Third, there are positions that require empathy and creativity in their fulfilment. This category includes jobs like teacher, nanny and doctor, but is much larger in size than the other job categories. These jobs require compassion, trust and empathy from the entities that act in order to fulfil the tasks required, which qualities AI does not possess. And even if an AI tried to fake it, nobody would prefer a chatbot telling them they have cancer, or a robot to babysit their children, Dr. Lee writes.
Though robots might not deliver the news of a health diagnosis to patients, AI is already being used to augment the work of doctors. For example, a team of Standford University scientists used AI to determine when patients will die in order to improve access to palliative care, or to specialized care for patients who have serious illnesses.
Fourth, there are the “As-yet-unknown’ jobs, which might become necessary in order to monitor and coordinate AI-based machine, and robotic technologies in workplaces.
For example, in the future, trucks, taxi’s and cars will be able to drive themselves, without needing individual drivers on the steering wheel, which creates new human jobs as fleet operators.
So, what will happen with “human” jobs, because of the advanced AI-based technologies?
Yet, AI-experts are convinced that future artificial intelligence (AI) and machine-learning (ML) technologies will focus on providing guidance in the necessary job-related tasks that need to be transformed for jobs in a variety of industries through technologically-advanced automation processes, and partial AI and ML takeovers of job-related tasks that were performed by humans before, instead of focusing on full automation of all job-related tasks that were originally performed by human workers.
Therefore, technology experts expect AI to transform how job-related task-activities will be performed in the future, which people will perform these task-activities, and how the time allocated to these task-activities is managed. Overall, AI and ML technologies may potentially influence job-related task-activities through complementations, and enhancements of human abilities in performing job-related task-activities, instead of performing all human-performed task-activities in people’s jobs.
Artificial Intelligence (AI) and “human jobs”
AI has been used by many businesses to automate job-related task-activities, but those that deploy it primarily to drive out workers will see only short-term efficiency gains. In an article that appeared in the 2018 issue of Harvard Business Review, a study was performed, including 1,500 businesses, which found clear evidence on great performance gains that were achieved by businesses that stimulated collaborations between humans and machines in performing job-related tasks. Therefore, I propose that humans and AI work the best together through collaborative intelligence to build on each other’s strengths.
|Human abilities||Artificial Intelligence abilities|
|Leadership||Speed of tasks-performed|
|Teamwork||Scalability of tasks-performed|
On the one hand, there are human abilities that cannot easily be captured by preprogrammed abilities of artificial intelligence mechanisms can be observed in AI’s difficulties with making jokes. The human ability of making jokes, being a type of behaviour that should be rather easy to exhibit for humans, seems to remain incredibly hard for AI-mechanisms.
On the other hand, there are abilities of artificial intelligence that cannot easily be captured by human abilities, such as AI’s rapid computer processes that have the ability of performing extensive data analyses on gigabytes of data, which would be just about impossible for humans to match. Therefore, it is not surprising that businesses and institutions are increasingly using collaborative intelligence of human abilities as well as the abilities of AI.
How can human intelligence work with artificial intelligence to produce augmented intelligence?
To answer this critical question, it is important to know how humans and machines can work together to make augmented intelligence. Garry Kasparov, a world-class chess player, has proposed a couple of new ideas on the topic of augmented intelligence. After having lost to IBM’s Deep Blue robot, Kasparov tried to figure out how cooperating with an artificially intelligent assistant could change high-level chess games.
Kasparov found that having a good process made a lot more sense. According to Kasparov, a Weak human + machine + better process, was always better than a strong computer alone, and even better, better than a strong human + machine + bad process. In the future, when leaders think about how to use AI-based technologies in their businesses, they will have to keep people’s expectations in check, will have to invest in teamwork and process improvements, and will need to work on the development of leadership skills in the existing group of “human” workers.
According to the financial experts of PricewaterhouseCoopers, artificial intelligence (AI) could add as much as 15.7 trillion U.S. Dollars to the world’s economy by 2030. The reason for this unusual increase in added value, is the increasing importance of data, and artificial intelligence in the way businesses make money, and compete with their competitors. The way that companies work is going to change because of AI. Many think that the people who do this work will also change. They think that organizations will start to replace human employees with intelligent machines. This is already happening: intelligent systems are taking over jobs in manufacturing, service delivery, recruitment, and the financial industry, moving humans to lower-paying jobs or making them unemployed, so this is already happening. This trend has led some people to think that our workforce may look very different in 2040.
People and machines don’t seem to be competing with each other, do they?
Machines nowadays seem to be able to do the work of our minds, as well as our bodies. With this new trend, it looks like there will soon be nothing that can’t be automated. But I think this view of how AI will be used in the workplace is wrong.
People vs. machines: Which of the two is smarter?
Artificial intelligence (AI) is a computer that acts and decides in ways that look like they are made by a person. In line with Alan Turing’s ideas, AI acts, feels, speaks, and makes decisions like humans do. It can be used to do lower-level tasks that are repetitive inside a closed management system. In an open management system, workers have to deal with outside pressures and changes. In this kind of environment, workers must be able to anticipate and deal with things like sudden changes in work environments. This is a different kind of AI, if you want to call it that.
When it comes to meeting performance requirements, which skills are most important?
When it comes to meeting performance requirements, I suggest that when artificial intelligence and authentic human intelligence work together, they will make intelligent work, the norm in the future of business and institutions. It will make organizations more efficient and accurate, but also more creative and proactive, because it will make them smarter.
Overall, I conclude that Artificial intelligence (AI) is not a bad thing for society and humanity in general. In line with this, I recommend businesses to invest in AI-based technologically advanced machinery taking over parts of former “human” jobs that are neither interesting for, nor worthy of being performed by, human workers who possess the socioemotional skills for higher level jobs that may not be as easily taken over by AI-based technologies as routine, quantitative task-activities that are part of one’s job. In order to get to that level of intelligently using AI in business, it is important for the people in political or institutional leadership roles to figure out how artificial intelligence might affect the ways how humans think and act in the workplace. There are several things that will have to be changed to the way in which human teams interact with AI-based technologies, as their newly developed member of the team. In order to smoothen this process of adaptation and development, it is very important for political, institutional, and corporate actors that fulfil leadership roles to improve their coordination, and coaching skills to match with the future needs of these technologically-advanced team processes. Human authentic intelligence, and artificially intelligent machines will have to work together more than is done right now. This means that businesses and institutions need to expand their workforces, whilst also improving their efficiency levels at the current human-filled jobs. Therefore, I recommend businesses to not prefer one option over another option, but to balance the use of artificial intelligence and authentic intelligence in the workplace, in order to achieve great results by making use of the improving capabilities of augmented intelligence.