Automation and AI are often perceived by companies that leverage them as an important source of labor productivity. Many workers in such companies, however, tend to see the adoption of these and other technologies as putting their jobs in jeopardy or creating more stressful workplaces.
Recent research highlights the dichotomy. While increased robot use contributed approximately 0.4 percentage points to annual labor productivity growth in major developed countries from 1980 to 2014, every additional robot per thousand workers that was deployed in the same period reduced the employment-to-population ratio by about 0.2–0.3 percentage points and wages by 0.3%–0.5%. While information and communications technology in the same period and same countries contributed to one-third of total economic growth, technology diffusion in enterprises has contributed in likely the same proportion to increased worker stress.
For companies eager to invest in AI and automation, the fears around adoption could be material, with heightened risk aversion creating unintended negative consequences. What if workers fear the future so much that it changes their behavior as consumers and leads them to spend less? If stress levels rise to such a high level as workers interact with new smart machines, could labor productivity suffer significantly, even going so far as to eliminate the benefit of automation and workplace changes?
Our latest research delves into these issues by analyzing the effect of AI diffusion on various drivers of welfare that include but also go beyond the usual GDP indicators. Alongside income per capita growth, these factors include: the risk of an increase in income disparity, as technology boosts workers with high cognitive skills and substitutes for those with low cognitive skills; the risk of unemployment, driven at least temporarily by a skills mismatch and limited inter-industry mobility; the higher intensity of jobs leading to possible stress; and improved longevity as a result of AI breakthroughs in disease prevention and delivery of more effective and personalized medicine.
Our data simulations show that the outcome for business, the economy, and societal welfare more broadly could largely depend on two critical factors. The first is the extent to which AI is deployed to create significant product and services innovations rather than primarily to cut costs and substitute for labor. The second determining factor is how AI is used to manage and smooth the inevitable frictions that will arise during AI-based economic transitions.
In general we found that fear about the risk of unemployment and increased income inequality could be enough to reduce citizens’ welfare. And if the risk aversion were sufficiently large, this would counterbalance the benefits of higher productivity and income to be had from the deployment and diffusion of the technologies.
On the other hand, the dual approach of focusing deployment on innovation and proactively managing the transition, through retraining and other steps to enhance worker mobility, could have a significant upside. Doing good for society in the AI era, according to our analysis, implies that implementing such a positive technology strategy can be good business.
For students of history, this may not come as a surprise. Henry Ford famously realized that his workers would be the first customers of the Model T and started paying them $5 per day in 1914, twice the typical daily rate, even as he cut prices for the cars by 50% over five years. The result was a significant increase in his company’s productivity, profits, and employment.
A century later, our research suggests the same can hold true today: Companies that manage the risks of a technology transition by focusing on welfare and well-being can benefit from the outcome, even as they become more competitive from their AI investments. There are several ways for companies to achieve this. They include:
Energizing workplace well-being through AI. An important (and often overlooked) point linked to technology diffusion is how technology has improved health and longevity over the decades. AI’s potential to improve diagnoses of, treat, and even cure chronic diseases is well documented — and holds significant promise for mankind. At a less lofty level, AI may also be used to reduce stress and other work environment insecurities. Stress, work safety, and fears about jobs are often cited as the largest sources affecting organizations’ productivity today. As a broad base of the workforce is affected, companies with the most effective health programs can experience significantly better financial outcomes.
What can companies investing in AI do with the technologies to better support well-being? Many companies are deliberately using AI not to kill jobs — and they communicate about it. Instead, they are focusing their use cases on reducing work overload, including rote tasks in call centers, or expanding business in a way that adds to employment, for example through new product innovation or by entering new markets. Companies use sensors and AI and analytics for predictive maintenance, and this also helps avoid the risk of human accidents and injuries. Some companies use AI as a tool to improve employees’ well-being — and see a strong return on investment on such spending. Johnson & Johnson has a long track record of focusing on employee health. It started its first wellness program, called Live for Life, in the late 1970s and continues to expand its program with the goal of having more than 100,000 employees at their well-being best. To that end, the company has convinced more than 90% of employees to use a health app that uses AI tools to personalize training and actions during work as a way to improve on health markers such as high blood pressure or cholesterol.
Other tech applications can increase happiness and employee engagement. For example, Vibe is an algorithm that analyzes keywords and emojis sent among employees on Slack to gauge whether a team is feeling happy, stressed, disappointed, or irritated. Communication analysis AI tools such as ADP’s Compass give managers insights on employee morale — and offer subtle nudges on how they could boost it; ADP reported that its management effectiveness improved almost 40% after implementing the technology. Humu uses data analytics to identify behavioral changes that are likely to make the biggest impact on raising the happiness level of workers. Then it uses emails and text messages to nudge individual employees into small actions that advance the larger goal.
Improving workers’ skills using AI and co-bots. Skills that will be needed to thrive in the workplace of the future are evolving. Digital skills, higher cognitive skills, and social and emotional skills will all be in greater demand, according to our prior research. AT&T has invested $1 billion in a sweeping program to train its workforce in digital skills. Walmart has also been investing large amounts, up to $2.7 billion since 2015, in training programs such its Walmart Academy and its Pathways program. It has brought into stores AI-based co-bots that “train” retail workers in how to collaborate with them. The bots have taken over mundane shelf-scanning and floor-cleaning duties, freeing human employees to spend more time assisting customers. AI technology can bring other elements of reskilling — platforms such as Slack play a key enabling role in creating a new workplace for crowdsourcing innovation and ideas, bringing new cooperation and creative skills, and preserving and upgrading employee talents, all while boosting returns on innovation.
Boosting salaries and incentives. Reminiscent of the Industrial Revolution and Henry Ford’s material increase in wages for his employees at the start of the 20th century, Amazon recently raised its minimum wage by a material amount, and Facebook has also raised its wages. In general, wage increases can be a strong signal. Just as Ford viewed higher salaries as a way to boost demand for cars, so higher wages and training may encourage better-earning employees to buy more from Amazon in the medium term.
Be more proactive in local communities. Some companies are focusing on spurring greater involvement with local communities, including through public-private partnerships. Gap, for example, provides job skills to disadvantaged people. Google’s “Grow with Google” initiative gives access to training and tools to different communities including military spouses, small business owners, teachers, and startups.
These are just a few examples of the types of measures companies could consider as they seek to capture the benefits of automation and AI at the same time as enhancing well-being. What is important is that the two aspects go hand in hand. The upsides for corporate performance and for society’s welfare work in tandem, according to our research. For business leaders, finding an approach that fits a new mold of technological social responsibility will be one of the key challenges of this era of AI and automation.
This article first appeared in Harvard Business Review.