“Tech Support”, Cover of The New Yorker October 23, 2017.

Source: R. Kikuo Johnson

Artificial intelligence or AI was defined as early as 1956 by John McCarthy, a Dartmouth College professor like this: “Every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it.”

AI is everywhere and human beings have paid attention to AI for several decades – but things reached a new high point after Google DeepMind’s success at the game of Go, an abstract strategy board game.

In May of this year, AlphaGo, an AI program that can play Go and was developed by DeepMind, defeated the world’s best human player, China’s Ke Jie, by 3 – 0, which demonstrated for the first time that an AI has surpassed humans in this game.

Most people already know that IBM’s Deep Blue defeated chess grandmaster Garry Kasparov in 1997 and after that, AI programs have defeated humans in several other cerebral contests. But “AlphaGo’s success is considered the most significant yet for AI,” said Agence France in an article on theguardian.com, as the game of Go is very complex – it is estimated that the game of Go has around 10170 possible configurations for each match.

These striking examples of AI defeating humans on games made people around the world nervous about their futures. But when asked in surveys, people have different opinions about AI, the potential drawbacks and benefits AI will bring to society, and if jobs are threatened by AI.

Overall feelings about AI

It’s no doubt that a human like Ke Jie has still more brainpower than a computer. An AI brain like AlphaGo might outperform a human being at just a few things. But the fact is that AI is becoming smarter than people at most tasks, which drives public debates about AI’s potential impacts on employment.

AI’s potential impacts on jobs

A 2017 survey conducted by Northstar Research Partners Ltd. among 3,938 global participants showed that globally, the biggest issue for respondents is AI’s potential impact on jobs, with job-related concerns (30%) seen as the biggest drawback of a future in which AI impacts human life.

Results from the Northstar survey also showed that participants recognize the profound impact AI may have on current society with the biggest potential benefit perceived to be medical and scientific advancements. Indeed, medical AI applications are growing rapidly. Other benefits include the replacement of tedious and dangerous tasks by robots or intelligent machines (29%), being able to save business money and increasing capability (18%), keeping humans away from danger or accidents (11%) and making it possible for human beings to have more free time (5%).

Also, according to a 2015 survey conducted by the Pew Research Center, among 2,001 adults, two-thirds of Americans think AI will do much of the work currently done by humans within five decades. Most people who already have a full-time or part-time job think that their jobs or professions will still exist in 50 years. Showing much confidence, it seems that they haven’t realize to what extent AI technology will encroach on human employment.

A 2013 study by researchers at Oxford University posited that as many as 47% of all jobs in the United States are at risk of “computerization.”

The research shows that there are high probability occupations likely to be substituted by computers programs and AI relatively soon.

Why are jobs under threat?

Over the past decades, computers have become substitutes for various types of jobs for many reasons, such as recent developments in AI machine learning (or ML).

ML will also reduce great demand for labor input in tasks because they can be routinized depending on pattern recognition, while increasing the demand for labor-performing tasks that are not subject to computerization.

Research done by Carl Benedikt Frey and Michael A. Osborne has a couple of interesting findings. First, most labor in these occupations will be replaced by computers: transportation and logistics, office and administrative, and production.

Transportation and logistics occupations

There will be more and more vehicles with advanced sensors that are increasingly cost-effective thanks to the development of computerized cars and the declining costs of sensors.

Office and administrative occupations

Office and administrative occupations are also at high risk of being replaced because algorithms for big data are rapidly entering domains depending on storing and examining information.

Production occupations

ML will reduce much demand for routine tasks that industrial robots are now taking on most operations in manufacturing. Industrial robots are becoming more advanced with their increasingly developed senses and dexterity, and they have greater potential for performing a wider scope of non-routine tasks. From a technological point of view, even though there are still many people who remain employed in production occupations, their jobs will diminish over the next decades.

A more surprising finding of the research is that a lot of jobs in services, sales and construction exhibit high probabilities of computerization.


Annually, there has been about 20 percent growth in the market of personal and household service robots, according to MGI, an international database website. Indeed, human beings have comparative advantages in tasks requiring mobility and dexterity but the advantages are decreasing, which will speed up the pace of labor substitution in this occupation even further.


High risk sales occupations include cashiers (97% are likely to be at risk), counter and rental clerks (also 97%), telemarketers (99%, which makes it the most susceptible job to computerization), and so on. These occupations indeed involve interactive tasks, but they do not necessarily require a high degree of social intelligence.


Prefabrication enables construction work to be performed under controlled conditions in factories, which makes tasks less variable.

AI can’t do it all

Which jobs are hard for AI to do well? Carl Benedikt Frey and Michael A. Osborne said in their research project that “Occupations requiring knowledge of human heuristics, and specialist occupations involving the development of novel ideas and artifacts will be difficult for AI to replace.”

With technological improvements, the comparative advantages of humans’ capacity for perception and manipulation of tasks could eventually disappear. This will require the ability to restructure tasks in innovative ways.

Carl Benedikt Frey and Michael A. Osborne also said that If jobs or tasks involve aspects of fine arts, originality, negotiation, persuasion, social perceptiveness, and assisting and caring for others, they will be in the low risk category(the probability is less than 0.3).

Predictions are that most management, business, and finance occupations, which are general in nature and require social intelligence, are largely confined to the low risk category.

The reason why engineering and science occupations are not highly subject to computerization, on the other hand, is largely due to the high degree of creative intelligence they require.

Actually computers are perfect complements to humans in creative science and engineering occupations, although it is possible that computers will fully replace workers in these occupations in the long-run.

Where do we go from here?

Indeed, AI will have an impact on jobs but It’s hard to predict how much and when. The question is: Will AI’s impact be positive or negative?

The analysis in the survey conducted by Northstar shows that on one hand, advanced AI technology can increase efficiency and productivity, guarantee the safety of employee, and even a better living (more free time, lower prices). But AI may also cause a large number of jobs disappear and make a significant change to traditional employment. At the same time, new jobs and sectors will then take their place.

So what comes after computerization? Bernard Marr call it the fourth industrial revolution, or industry 4.0 in his book Data Strategy, representing the combination of cyber-physical systems, the Internet of Things, and the Internet of Systems.

Bernard Marr also said in his book that business leaders have to widen their visions from tradition in order to thrive, and begin to question everything, from reconsidering their strategies and business models, to discovering the right investments in training and potentially disruptive R&D(research and development) investments.

The future is ours to shape. People, businesses, and nations must get ready for it and thrive in the new, smart global economy.