The future of work and the work we must do now: Davos reflections

by James Manyika, Chairman of the McKinsey Global Institute

At this year’s World Economic Forum in Davos, many of the conversations I had focused on artificial intelligence (AI)and automation. Recent research at the McKinsey Global Institute (MGI) has touched on the economic benefits that can come from AI and automation adoption, particularly given weak recent productivity growth. Other MGI research shows some of the promising ways that AI, while not a silver bullet, can contribute to efforts to tackle the world’s biggest challenges including the United Nations’ Sustainable Development Goals. Still, my colleagues and I led discussions about our concerns — and the concerns we heard from business leaders and policy makers — on the coming disruptions for work and jobs, and what to do about it to ensure positive outcomes for workers and society. This is a topic MGI has been studying deeply over the past few years and collaborating with others on, and one we think is hugely important to tackle if technology is to make our lives better.

The most common concern when thinking about AI and automation is that we may face a “jobless future” as a result of these technologies. Our research suggests that is unlikely to be the case in the next decade or two. That is because economic and productivity growth and other catalysts driving demand for work are likely to create more jobs than automation may displace, at least on a net basis and viewed at a national level.

Let me qualify this with some important nuances. First, while this holds true in most of the scenarios we considered, some extreme tech acceleration, adoption, and economic stagnation scenarios are also possible, and these could lead to a different outcome. Second, even if jobs gained outnumber jobs lost overall, there are likely geographic variations and challenges at the county, region, or state levels — for example the jobs may not be where the people are or match the skills available. Third, for many developing countries, the jobs challenge is primarily driven by demographics and the pressing need to create large numbers of jobs to match the number of young people entering the workforce.

Automation — along with other digital technologies — makes meeting this challenge harder in two ways: first, the external markets for whom these developing countries historically may have served as sources of labor are likely to need less of it as global value chains evolve, and second, some local activities themselves may be automated, though likely fewer than in advanced economies, given wage structures.

Even in a world with enough work overall, our research seeks to draw additional attention to pressing concerns about the coming workforce transitions as a result of automation:

• First, as some occupations with a higher density of automatable activities decline (jobs lost) while others grow (jobs gained), millions of workers may need to change occupational categories. These occupational transitions (many involving locational transitions) are already underway.
• Second, workers’ skills will need to shift as employers seek more technological and social and emotional skills, with less demand for physical and basic cognitive skills, though these will continue to dominate for some time.
• Third, there could be pressure on wages, given the occupation mix shift; many occupations that may decline have tended to have middle-wages while many of the growing occupations have historically paid less. The other concern on wages is around partial automation. While there will be many cases where humans and machines each perform value-added work and wages rise, as seems likely for occupations such as radiologists, there may be cases where the value-added portion of a job may be automated, leaving the less skilled work for humans.
• Finally, workplaces and workflows will need to be redesigned to better accommodate human-machine collaboration.

One of the reasons we remain concerned about these transitions is that we start in a hole on most of them — in this sense, the future of work is already here and requires action now. On occupational transitions, most advanced economies did not support workers well in making the occupational transitions during the last wave of globalization — indeed most OECD countries have been reducing the scale of mechanisms for supporting worker dislocations and transitions. On skills, in survey after survey companies highlight the shortage of skilled workers. Moreover, at a time when more workers will need to learn new skills or raise the level of the skills they have, already low spending on worker training has declined as a percentage of GDP over the past two decades across many countries, including the United States. On wages, income inequality and wage polarization have been a significant issue in recent years; we have already seen continued stagnation and slow wage growth, worsened by rising costs of living particularly in healthcare, affordable housing, and education. This wage stagnation issue is one we studied in detail two years ago in our report Poorer than their parents. And all this is before the bulk of adoption of automation technologies seems set to occur.

At the same time, I do want to highlight areas where we are more heartened. We have seen some progress, particularly on the skills and reemployment front, through some organizations we have partnered with. They include the Markle Foundation where I am involved, whose Skillful Initiative is now working across 20 states. I have also been heartened by some employers who are taking the lead on skills, such as Google.org’s future of work portfolio, and Walmart, SAP, and AT&T’s reskilling initiatives, as well as McKinsey’s own work on Generation. Walmart’s Kathleen McLaughlin has discussed the scale of their programs: 200 Walmart Academies across the country have trained 325,000 workers in the past few years. AT&T is working to reskill 100,000 workers, or nearly half their workforce. Elsewhere, in Germany for example, my colleagues are collaborating with a consortium called the Stifterverband, made up of business leaders across all industries and leaders in the higher education system to tackle reskilling challenges at scale.

There are also many others who are working to constructively highlight these workforce issues. One example is a forthcoming film Cyberwork and the American Dream by James Shelly (director) and Professor Elizabeth Cobbs (writer) that will air nationally on PBS in April, with which we were involved. Another example I am involved in is the Aspen Institute Future of Work Initiative, a bipartisan effort that brings together business, policy, and labor organizations to develop solutions and policy recommendations that address the challenges American workers and businesses face due to the changing nature of work in the 21st century.

Much more still needs to be done — and that’s going to require everyone’s involvement, including the private sector and policy leaders. In a paper last summer, my colleague Kevin Sneader and I laid out the ten key areas to solve for, and at scale, in order to lead toward a better outcome on the future of work, from economic growth to supporting workers, to investments that drive demand for work. One imperative that needs a deeper discussion will be to ensure we adopt these technologies safely, given the need to protect data security and privacy, as well as guard against bias and malicious use. The magnitude of the challenges we face, and the difficulty of our starting position, only demonstrate the continuing need for action.

This article was originally published on LinkedIn.