You are here: American University School of International Service Centers Security, Innovation, and New Technology AI and the Future of Work in the United States

Technology

AI and the Future of Work in the United States

By  | 

Artificial intelligence will change the future of work in the United States, but the consequences will differ across regions and industries. As with past advances in automation, AI will lead to increased levels of productivity, specialization in job roles, and an increased importance of “human skills'' like creativity, problem solving, and quantitative skills. Although AI will increase economic growth, these gains will not be evenly distributed. Rural communities that already face high levels of job insecurity will come under additional strain. AI will benefit labor in some industries but threaten it in others. Automation will complement job roles in high-growth fields like healthcare, where there is no substitute for highly-skilled practitioners, but replace jobs in industries relying on standard routines. To address these gaps, policymakers should focus on restructuring school curricula to reflect the changing skill demands, addressing the educational attainment gap between rural and urban adults, and providing economic relief to workers forced to leave the workplace and learn new skills.

Introduction

The evolution of technology and changes in the workplace have always happened in tandem. Beginning in the 1760s, a series of industrial revolutions ushered in mechanical innovations like railroads and then mass production factories using man made tools to push past the limits of human manual labor. Then in the twentieth century, computers began to replace human cognitive work. Optimists viewed this as a sign of inevitable progress toward a more productive society, with lower prices for goods and increased opportunity for specialized labor; but there has been a pervasive concern that machines will replace human labor and lower the number of new jobs.

With the rise of personal computers, automated call centers, and industrial machinery during the early twenty-first century came a massive net increase in employment with 54 million new jobs. However, a not-so-subtle exacerbation of inequality in America accompanied this explosion in efficient production. A widening wealth gap, bottoming out of the middle-class, extreme job-loss in rural communities, and stratification of education and job opportunity have become commonplace. The advent of digital technologies was particularly difficult for manufacturing and clerical workers. A seamless transition into a new job market did not occur for the newly unemployed; rather, those who lost their jobs during the rise of automation were crammed into service industry jobs that paid less than their previous occupations had. The disparity that has risen during this period is unlikely to reverse itself. Indeed, it will continue to worsen as technology outpaces the capabilities of the humans who built it.

More recent innovations in computer science have rekindled fears of more human displacement and job loss. Machine Learning and narrow Artificial Intelligence (AI) have become ubiquitous, and unlike previous waves of technological change, these innovations are affecting most, if not all, industries. Now the past must serve as a guide as to how to prepare for future job insecurities, especially by investing in education, skill retraining, and renewal of rural communities,where the opportunity gap threatens to worsen job displacement in the face of AI-based innovation.

The Basics of Narrow AI and Machine Learning

        Artificial Intelligence is often conflated with dystopian science fiction, as in killer robots from The Terminator or hallucinatory computer programs from The Matrix. But the technology-centered human experience of today is far more nuanced than sensationalized fictitious depictions. Various forms of artificial intelligence are commonplace; today, machine learning powers everything from voice-activated “virtual assistants” on most smartphones, to recognition systems built to improve accuracy in medical diagnoses. It also increases the scope of industrial automation, which has raised concerns about the effects on labor.

The fundamental concept that powers machine learning is the notion that a computer can learn from new information. Designed in a loosely similar fashion to the human brain, some forms of artificial intelligence rely on constructing an artificial neural network to build connections that replicate the complex level of understanding characteristic of humans. Once an algorithm is built for a machine, it will undergo three different forms of learning, supervised learning, self-supervised learning, and reinforced learning, which develop and strengthen its neural network. Once the computer understands its intended function, it works independently to self-correct through a series of trials and errors - a process that ultimately can lead to complex decision making.

The logic behind deep learning and neural networks is simple but the result is complex. One form of machine learning, often referred to as hierarchical learning, is accomplished using a large set of layered images or data points that are pieced together from the bottom up by the machine until it deeply understands a concept. For example, if given an input in the form of a matrix of pixels, the computer will begin by classifying individual pixels, then forming clusters of pixels, before finally identifying the full visual. This straightforward process is repeated tens-of-thousands of times with an equal number of new data sets. It results in a web of nodes that strengthen the predictive decision-making process of the computer.

This world is increasingly driven by the currency of data, and AI will undoubtedly have a deep, divisive impact on the jobs of tomorrow; many service-focused jobs in the food industry will face high susceptibility to automation, for example, as opposed to industries that will come to use it as a supplementary tool, like the medical field.

The Promise of Machine Learning and Narrow AI in Healthcare         

Unlike the physical human brain, an AI program can quickly discover problems in narrow, data-based situations, a time-saving benefit that already aids decision making on treatment options in healthcare, for example. Health care professionals use algorithms to improve diagnoses through a form of machine learning called precision medicine; this simple algorithm relies on previous patient information (and supervised learning) to assess what treatment options might be most successful.

Another useful application of narrow AI uses more complex deep learning algorithms to identify potentially cancerous lesions in radiology pictures, a process that aids in early diagnosis and can prevent the spread of malignant tumors. This tool allows professionals to allocate their own time more efficiently towards specialized tasks, such as devising creative treatment solutions and researching other medical problems, potentially leading to job creation.

Tech giants like Google are collaborating with health firms to build programming that compiles user data to identify patients at risk for conditions that are typically genetically predisposed but difficult to predict--like cancer. However, simple AI tools also extend beyond more precise diagnosis in the medical field: doctors use natural language processing algorithms for transcribing patient-doctor interactions, surgeons use robots to improve the precision of surgeries, and hospitals employ algorithm-driven processes to replace rote tasks within hospitals. The proper use of these tools in healthcare will change the way medicine is administered: it can improve patient care, reduce fatalities with early diagnoses, and create more jobs through specialization.

The Jobs of the Future

 Most jobs sectors that are predicted to be high-growth over the next decade will require high levels of education. One source claims that high-paying jobs requiring a bachelor's degree or above will face only 29% of job change due to automation. Like the field of healthcare, other “high-value” and professional, scientific, and technical service-oriented industries are predicted to experience low levels of automation (34%) accompanied by relative job growth (3.8%). Additionally, given the benefits AI poses to complex, high-skilled vocations, it is likely that AI-related jobs could see an increased share in the economy, a trend that would create more stable, well-paying positions for Americans. The assignment of rote tasks to machines and complex tasks to humans will increase productivity and enhance economic growth. One economist predicts that AI-related changes will result in an increased GDP of $3.7 trillion for North America before 2030. Unfortunately, these economic benefits will not be evenly distributed, and rural communities will absorb the greatest costs.

New technologies have sometimes had transformative impacts on the labor market; in this case, AI will have a striking effect on many service and white-collar jobs that have been relatively safe from automation to this point. Compared to the medical field, most routine, service-oriented roles will be phased out by the increased presence of AI. Jobs that are at the highest risk of automation include office support jobs and food service jobs. These jobs are ubiquitous but rural workers rely on them more since they make up a larger share of the rural economy. For example, the switch to automated stores has already started with AmazonGo, the prototype for an entirely contactless grocery shopping experience.

Service industries that will face the most severe levels of automation include the food service industry and office support roles. These jobs are generally held by individuals who haven’t earned college degrees with more than 90% of food service employees and more than 60% of office support workers having at most a high school diploma.

 There is a severe digital gap between rural and urban areas: 24% of rural adults experience issues with high-speed internet, whereas only 13% of urban adults say it is an issue. Substantial portions of rural areas still lack the infrastructure required to implement high speed internet, which makes it exceedingly difficult to bring high-growth jobs to these areas. There is little migration between cities, which suggests labor market fluidity. To make matters worse, businesses are investing less in their workers: the percentage of employer-sponsored job-training fell by 7 percent between 2003 and 2013 even though the need for high-skilled workers is rising.

Policy Implications

Although the situation may seem dire for low-skilled rural workers, new technologies might also provide some relief. Data compiled by the Economic Modeling Specialists International can be used by local communities to help pinpoint careers for displaced workers to transition into given their acquired skills from previous occupations. This along with strong federal education and economic policies could reduce the shock of a changing job market.

        At the top of the agenda is improving the American educational system, especially measures to rectify underperformance in STEM-related subjects. The most recent PISA test scores show that the United States ranks 36th out of 79 countries for math proficiency and 13th for reading proficiency. Given that future careers will be driven by a demand for digital skills in high-growth industries like data science and computer science, school curricula should reflect evolving employer expectations, by introducing computer coding and statistics, for example.

Also, high-paying and high-skilled jobs are concentrated in large coastal cities, but there is little migration between rural towns and urban metropolises. College completion rate gaps continue to widen between urban and rural areas with a 14% difference in the proportion of adults with bachelor’s degrees. Worker incentives to complete training and move to areas ripe with opportunities are important national priorities. The great challenge to bridge would be the existing inequality in educational attainment, and therefore, job opportunity--an issue that has fed into the discontent and general unrest felt in some rural parts of the United States.

In the interim, economic policy recommendations by prominent economists like Laura Tyson argue that a higher minimum wage (productivity has increased while the median wage has not), accompanied by a negative income tax which is paid for by a progressive consumption tax can balance the wage inequality caused by increased automation and subsequent job loss. Additionally, there should be an agenda to increase economic mobility of workers in rural areas Other economists propose a Worker Training Tax Credit which is modeled after the Research and Development Tax Credit to incentivize companies to invest into training their low and middle-income employees. This tax credit would pay for training for employees earning less than $120,000 per year while encouraging business innovation.

Automation through AI is already occurring, and the pace of innovation demands organized collaboration towards regulation. In terms of labor, technology has always been both an asset and a liability, and evolving forms of artificial intelligence are no different. Short-term policy fixes will alleviate immediate suffering for those in the workforce displaced by machine learning and narrow AI. We must have long-term U.S. federal policy changes, especially those that incentivize education and re-training, if we are to manage the dramatic labor changes underway. 

 

 


About the Author: 

Pragya Jain is a 2020-2021 CSINT Fellow and current undergraduate sophomore studying International Relations with a minor in Data Science. She has an interest in how emerging technologies can greatly change global relations. Her research focuses include clean energy and technology as well as the implications artificial intelligence will have on humanity.


*THE VIEWS EXPRESSED HERE ARE STRICTLY THOSE OF THE AUTHOR AND DO NOT NECESSARILY REPRESENT THOSE OF THE CENTER OR ANY OTHER PERSON OR ENTITY AT AMERICAN UNIVERSITY.