domeet webmaster

For many people, the prospects for work seem to be getting bleak. According to a recent study by Forest Research, it is estimated that 10% of jobs in the US will be automated this year. Another study by McKinsey & Company also shows that nearly half of jobs in the US may be automated in the next decade.

The main things that can be automated are some repetitive and routine work, ranging from reading X-rays (the radiologist’s role may be more limited), truck driving, to warehouse management. Although we have seen a lot of articles about the types of work that may be eliminated, there is another point that is rarely published, that is: don’t ask which jobs will be eliminated, but ask what the surviving work will have. Replaced by the machine.

For example, doctors, it is clear that the ability of the machine to diagnose diseases soon will exceed humans. When data sets can be used for training and testing, machine learning works very well, and they can diagnose a wide variety of diseases. However, imagine what it would be like to sit down with the patient and family to discuss the treatment plan. For the foreseeable future, this situation is much less likely to be automated.

Another example is a career: a barista. In San Francisco, all the baristas at the Cafe X coffee shop were replaced with industrial robots, which meant that the robot replaced the position of the barista. When making hot drinks, robots use their antics to entertain customers. Even so, another employee was hired in the Cafe X coffee shop. He is responsible for teaching customers how to order drinks and solve some of the problems that these machine barristers have.

Comparing the two professions of barista and bartender, customers often talk to bartenders more often. Obviously, the bartender’s job is not just mixing drinks together. Just like a doctor, it is easy to break down a doctor’s work into two parts: repeated routine work and work with patients. This is like the work of a coffee shop can be divided into making coffee and serving guests.

After considering a lot of work and industry characteristics, we found that two kinds of unconventional work seem to be particularly common, but it is difficult to automate.

The first is the emotional class. Emotion plays an important role in human communication. For example, doctors who communicate with patients and bartenders who interact with customers. These two kinds of work almost all involve non-verbalAll forms of speech and emotional communication. But more importantly, it inspires us to consider the priorities of what we are doing now, for example, it helps us decide what we need to do now, and it is too late to make decisions.

Emotions are not only complex and subtle, they also affect many of our decision-making processes. It turns out that scientific understanding of the role of emotions is very challenging (although progress has been made) and it is difficult to build into an automated system.

The second is the environment. When humans make decisions or interact with others, it is often easy to take into account the environmental factors they are in. The environment is really interesting because it is open. For example, whenever there is news coverage, it will change the environment we live in (this environment, regardless of size). In addition, changes in the environment can not only change the interaction between various influencing factors, but also introduce new influencing factors and fundamentally change the organization between various factors.

For machine learning, environmental change is a big problem because machine learning is about working with data sets that were created in advance in different environments. Therefore, environmental factors are also a challenge for automation.

We have the ability to manage and use emotions and the ability to take into account environmental impacts. These are key components of critical thinking, creative problem solving, effective communication, adaptive learning and correct judgment. It turns out that it is very difficult to write a program that allows the machine to completely mimic human knowledge and skills. It is unclear when or if mature technology will truly replace humans.

In fact, the talents with these two abilities are also sought by employers in all walks of life. In one survey, 93% of employers said that job seekers are more important than their undergraduate majors if they have critical thinking, good communication skills, and the ability to solve complex problems.

In addition, employers also want candidates to have other “soft skills,” such as strong learning skills, the ability to make the right decisions, and good collaboration with others. Of course, there is no problem with these popular abilities, but to achieve automation, it is difficult now and in the future.

All of this suggests that our education system should not only focus on how people interact with technology (for example, teaching students to program), but also teach them how to do things that machines can’t do, and look beyond the long-term. This is a new way to describe the nature of “soft skills.” These “soft skills” may be mistakenly considered to be incomprehensible and systematic skills, but they bring advantages to humans and will continue to give humans an advantage over robots in the future.

Stephen M. Kosslyn|文

Stephen M. Kosling is the Dean of Foundry College, the former Chief Academic Officer of the University of Minerva, and the former Head of the Department of Social Sciences at Harvard University.

阿丫丫|译周强|校