The AI paradox: How robots are making work more human

Over the next decade, a great wave of technological change will wash through the economy, transforming the nature of work and the shape of the labour market. In a recent study we carried out in partnership with Cisco, we built a brand-new modelling framework to explore the implications of this change in a more comprehensive way than ever before. We simulated the real-world dynamics of technological change and its interaction with the labour market and found that 6.5 million US workers will have to seek out a new profession over the next decade.

But what significance does that number have? Our bottom-up approach to understanding the evolution of the labour market around this technology trend enables us to identify what needs to be done to meet the needs of the future economy. After all, at the macroeconomic level, our ability to take advantage of the opportunities technology presents will be defined by how smoothly those 6.5 million workers make the transition from the old to the new.

Here are our seven key implications for the future of work:

  1. Technology causes a change in the nature of work, rather than the number of jobs. The people whose jobs are most vulnerable to disruption are those who do routine and repetitive tasks, whether cognitive (such as documenting information) or physical (such as moving things around). But automation will free these people up to work on tasks that technology cannot so capably, or cheaply, compete.
  2. Around 4.3 million workers will be displaced by technological innovation and forced to find new work. But a worker displaced from a security guard job is unlikely to fill a vacancy in web design. To simulate worker movement, we build a sophisticated skills-matching model that fills job openings iteratively across industries and up and down the skills hierarchy by finding candidates that most closely match the skills and experience required in the role, and best compete on price.
  3. As the labour market evolves around technological change, vacancies will cascade throughout the labour market. The displacement of 4.3 million workers will lead to almost 6.5 million total job moves before the future labour market returns to equilibrium. This is over and above the everyday “churn” that normally occurs in the labour market and is equivalent to 4.9% of the current working population. To focus on the displacement effect alone, misses a large part of the picture.
  4. The “hollowing out” phenomenon, in which the middle-skilled jobs in the labour market are squeezed as low-skilled and high-skilled work continues to grow, will continue into the future – although not accelerate. In fact, whilst a broad hollowing out trend holds in our forecast, the underlying data is much more nuanced. We see displacement taking place at the top of the skills spectrum and the bottom, as well as a growth in demand for jobs in the middle. The impact of technology will depend on the specific characteristics of each role, as well as the demand for the goods and services workers are producing.
  5. Workers in transport, lower-level manufacturing and agriculture jobs will face a difficult time in the next decade. These are the occupations for which the risk of automation is relatively high and the “income effect”, by which new demand for goods and services drives up the demand for workers, is less strong. Nonetheless, new jobs will be created in many areas, such as computing, management and media, as well as in healthcare and sales. For instance, we predict four percent more healthcare workers will be required in 2027 and almost seven percent more salespeople – a category including online business development as well as in-person retail assistants.
  6. Looking across the economy, the US is facing a significant reskilling challenge in the next decade: not least in meeting an acute ICT skills shortfall. Advanced and entry-level programming skills top the training requirements, with new workers bringing only around 40 percent of the skills they will need to new programming positions. This is driven by rising demand for specialists in the ICT sector as well as a proliferation of lower-level programming jobs in the arts, healthcare and financial services sector. The ICT skills shortfall also extends beyond programming to other associated skills like operations analysis, systems evaluation and technology design.
  7. But paradoxically, we also found that one third of the cumulative skill shortfall facing the US labour market in the next decade is in what we refer to as “human skills”. These include negotiation, persuasion, service orientation and social perceptiveness. Success in these areas will be defined by people’s ability to out-perform technology, or to work creatively with technology, rather than competing against. Those who are most adept at leveraging these inherently “human” skills will be rewarded. For traditional workers like truck drivers and machine operators, it is a shortfall in these “human skills” that may well define their ability to adapt to demands of the future economy.

We plan to explore the vast underlying dynamics to these trends, using the Oxford Economics Occupational Databank and Skills Matching Model over the course of the year.

Take a look at our report on the AI Paradox and read about our methodology in more detail, here.