Each wave of technological innovation brings fears that humans are destined for displacement from the workplace. Ever since the Luddites smashed machinery in the early 1800s, workers have protested, sometimes violently, against the threat of automation leading to job losses. Today, the increasing deployment of AI and machine learning has triggered similar concerns.

In some respects, these fears are well-founded. Despite current anxieties, in my new book, The Magic Conveyer Belt: Supply Chains, A.I. and the Future of Work, I argue that the remorseless advance of tech-driven innovation will not lead to the extinction of mass employment, especially if we’re smart enough to prepare workers for a different future. 

More on the Intersection of AI and WorkIs AI Coming for Your Job?

 

Does AI Mean Job Reductions? 

In many cases, automation’s impact on jobs typically follows three overlapping phases. The first phase is deskilling as tasks that previously required special expertise become partially automated. The jobs still require workers but with less mastery and, in some cases, lower pay. 

3 Phases of Automation-Related Job Disruption

  1. Deskilling
  2. Scaling
  3. Elimination

A modern example is London’s cabbies who, for many years, have been required to pass the toughest test in the world to prove they have “The Knowledge” of London’s byzantine maze of 25,000 streets and 100,000 landmarks and businesses. In 2012, however, Uber entered London with a navigation app that let anyone with a car and a smartphone ferry passengers around.

The second stage is scaling, which involves amplifying human labor so that each person can handle more work. This way, a given volume of production requires fewer workers. For example, the invention of telephone numbers meant that each switchboard operator could handle more calls without having to remember the switch setting for the requested name to make a call. The faster process allowed fewer operators to take a larger volume of calls.

In the third phase, elimination, the machines get better over time and can finally replace the deskilled jobs. Many examples illustrate this process. Computerized technologies, especially personal computers and office software applications, reduced the need for stenographers, typists, secretaries, and bookkeepers. Likewise, early passenger aircraft required a cockpit crew of five (pilot, copilot, navigator, flight engineer, and radio operator) but the progress of electronics eliminated these roles and left only the pilot and copilot with jobs.

The impact of elimination is, of course, job reduction. Some workers still need to manage tasks that are difficult to automate, however. These include personal service jobs, negotiation of business relationships, overseeing all the new machinery, and more.

 

Technology Gives As It Takes

New technologies can also spark job creation in four main areas.

4 Ways AI Can Spark Job Creation

  1. The creation of new occupations.
  2. Expanding current roles.
  3. Increased scale and scope.
  4. Other knock-on effects.

 

1. The creation of new occupations

The information technology revolution created many previously unimaginable occupations, such as software architects, data scientists, web developers, digital marketing specialists, and many others. Although the pace of change makes new opportunities difficult to predict, we can more confidently anticipate which jobs will likely be changed, deskilled and even eliminated. Such forecasting allows for people, companies and governments time to prepare for the changes through reskilling and changing careers. 

 

2. Expanding current roles

Some new jobs will be an extension of existing ones. As personal computers proliferated, spreadsheet software enabled ordinary workers to create their own numerical and financial models without requiring a programmer’s assistance at every step. This development did not eliminate central IT jobs, though, because workers still needed help, training, and software updates. 

Today, new generative AI tools will create demand for workers who can use them efficiently for increased productivity. Some of these jobs may include AI trainers and explainers so people can use the tools efficiently and realize when it gives erroneous results, customer service representatives that will focus on interactions regarding non-standard questions and offer personalized service, data scientists and ML engineers who can train the models and continue adapting it to changing environments and needs, and many others.

 

3. Increased scale and scope

Tech advances can generate more jobs. For example, reducing an airliner’s cockpit crew from five to two reduced the labor costs per passenger and encouraged more people to fly. In turn, the higher volume of passengers inspired scale innovations, including larger aircraft, further reducing the cost per passenger, which meant airlines needed to employ more pilots, cabin crew, baggage handlers, and airport workers.

 

4. Other knock-on effects

The increase in air travel caused international travel to boom; the number of international tourist arrivals worldwide increased by a factor of about nine between 1970 and 2019. As a result, employment surged in the travel and tourism industries. In 2020, almost 700 million people were working in travel and tourism jobs around the world. Similar processes encouraged globalization, which grew due to the adoption of containerization, increased air travel, and modern communications technology. 

Such knock-on effect are particularly difficult to predict. They involve new industries and new ways of performing work that do not exist yet.

 

Is the AI Revolution Different for Workers?

In previous industrial revolutions, some jobs disappeared. Each revolution brought many new types of infrastructure, however — railroads, electricity, roads, telephones, the internet — and new consumer products like cars, radios, TVs, appliances, computers, smartphones. These products, in turn, fueled innovations in retail such as department stores, mail order sales, shopping malls, big-box stores, and e-commerce that contributed to economic growth and more jobs.

The AI-driven revolution underway today is potentially different from previous versions in four main ways: it hits professional workers, affects a wide range of jobs, performs “uniquely human” functions, and moves fast. One of the problems in imagining the future is that we all know the jobs and people who are employed in today’s jobs who might be changed or replaced. They are our neighbors, co-workers, family members, and ourselves. We have little idea regarding the new economy and jobs that will be available in the future. Both of these facts are a source of anxiety.  

4 Ways the AI Industrial Revolution Is Different

  1. It hits professional workers.
  2. It affects a wide range of jobs.
  3. It performs “uniquely human” functions.
  4. It moves fast.

Current AI-infused technologies seem to be affecting jobs substantially further up the skill ladder than past industrial revolutions did. This, however, may be nothing more than another version of deskilling and a change in power dynamics within organizations. Just like easy-to-use computers and spreadsheets allow workers and managers to operate independent of central IT expertise, DALL∙E 2 allows employees to generate art without hiring an artist. Likewise, ChatGPT allows individuals to write software programs without the need for programmers. Over time, the increased productivity will likely lead to many new and different jobs.

Despite the claims that generative AI can act like a human, it is only a piece of software trained to work with many different elements (e.g., text, art, computer code, etc.). Given a query, the machine uses its logic, derived from its data, to generate a response. So, it will deskill some jobs but will likely lead to significant productivity growth, empowering workers adept at using the new tools, and create many new jobs.

The one factor that does make the current environment different is the pace of change of the AI revolution. The first three industrial revolutions took decades because they required mass production of new physical assets, such as replacing agricultural workers with machinery. Providers of those technologies had to design and build new factories (and new supply ecosystems) to make their products (e.g., power looms, steam engines, electrical equipment, container ships, etc.) at increasing scale. Workers had time to either age in place until retirement or slowly shift to other careers. Businesses and workers had time to adapt. 

In contrast, the transition to AI-driven automation, mobile apps, and cloud computing can be incredibly rapid. Although AI software may take time to develop initially, its mass production can be instantaneous and nearly free. Moreover, since most organizations have digitalized most of their communications, stored data, and business processes, switching those digital streams from human-mediated to AI-mediated systems can be swift. 

More in AIWhat Does It Mean to ‘Work With AI’?

 

Harnessing the Opportunities of AI

Although AI promises to usher in a new era of productivity growth, the speed of change is such that many workers may be displaced in the short run until new industries develop.  This can be a painful process, involving forced changes, reduced income, and other dislocations. Ignoring technological change is futile, however. Thus, countries, enterprises, and people need to prepare themselves for it. 

Even in a technology-based environment, workers and managers will need to understand the details of jobs. They are the ones who define the tasks for the machinery and, most importantly, supervise it and override it when necessary. The problem is that such workers have many years of experience. But in an environment in which some of the entry-level jobs have been automated, how does an organization create a pool of experienced mid-level workers if there are no entry-level jobs?

The challenge can be met in three ways. First, keeping up the skills of existing workers and upgrading them can tap into the huge and growing availability of online education and training. This includes both MOOCs and technologies that help people perform certain tasks at the time they’re needed, such as AI-based augmented reality, in which people are instructed how to perform the jobs in front of them.

Another solution for the U.S. is to adopt and customize apprenticeship schemes similar to the German duales Ausbildungssystem, meaning a dual vocational education and training system. This program is a three- to four-year program that combines about 70 percent time of work in a company with 30 percent training in a vocational school and provides salaries and subsidies for living expenses. More than half (54.5 percent) of German high school graduates go into this system, which covers 327 recognized occupations. In Germany, about two-thirds of the apprentices then accept a full-time job with the employer that originally hired them as an apprentice. Although the companies themselves offer these apprentices jobs — applicants apply to the company, not the school — the German government helps regulate and set educational standards for the programs.

Governments also need to plan for increasing unemployment funding, tied to retraining and moving people to new jobs as they emerge. The high rate of current technology development is such that the time is now to plan and develop new ways of dealing with the challenges of the coming changes in the work place and elsewhere.

The future of work is not devoid of humans. Instead, we must find ways to ensure that workers will have important roles in the new world that technology has always did and will undoubtedly continue to create.

Expert Contributors

Built In’s expert contributor network publishes thoughtful, solutions-oriented stories written by innovative tech professionals. It is the tech industry’s definitive destination for sharing compelling, first-person accounts of problem-solving on the road to innovation.

Learn More

Great Companies Need Great People. That's Where We Come In.

Recruit With Us