The Future of Employment


Paper examining how susceptible jobs are to computerisation. 47% of total US employment at risk, with wages and educational attainment predicting lowest probability of computerisation.


We examine how susceptible jobs are to computerisation. To assess this, we begin by implementing a novel methodology to estimate the probability of computerisation for 702 detailed occupations, using a Gaussian process classifier. Based on these estimates, we examine expected impacts of future computerisation on US labour market outcomes, with the primary objective of analysing the number of jobs at risk and the relationship between an occupation’s probability of computerisation, wages and educational attainment. According to our estimates, about 47% of total US employment is at risk. We further provide evidence that wages and educational attainment exhibit a strong negative relationship with an occupation’s probability of computerisation.


  • 47% of current US employment in the decade or two is at risk!
  • other papers already show that occupations mainly consisting of tasks following well-defined procedures that can easily be performed by sophisticated algorithms
  • workers reallo-cating their labour supply from middle-income manufacturing to low-income service occupations
    • service occupations require higher degree of flexibility
  • “As stressed by Schumpeter (1962), it was not the lack of inventive ideas that set the boundaries for economic development, but rather powerful social and economic interests promoting the technological status quo. … The balance between job conservation and technological progress therefore, to a large extent, reflects the balance of power in society, and how gains from technological progress are being distributed.”
  • transportation revolution, expanded market size, eroding local monopoly and increased competition
  • since electrification the 20th century is a race between education and technology
  • technological progress has two competing effects on employment:
    1. as technology substitutes for labour, there is a destruction effect, requiring workers to reallocate their labour supply
    2. the capitalisation effect, as more companies enter industries where productivity is relatively high, leading employment in those industries to expand
  • now higher skilled workers are being forced “down the occupational ladder”, lower skill out of labour force
  • training of Machine Learning (ML) restricted by data avilable for training
    • need enough to represent totality if input for training, plus at least half again to check for correctness
  • ML lacking some human biases
    • ex: experienced Israeli judges are substantially more generous in their rulings following a lunch break
  • ML works well for fraud detection and monitoring
  • in many judgement tasks the lack of biases is a comparative advantage as is the speed at which decisions can be made
    • ex: financial trading
    • advancements even in software development, bug detection
  • occupations that involve complex perception and manipulation tasks, creative intelligence tasks, and social intelligence tasks are unlikely to be substituted by computer capital over the next decade or two


  • it is possible that the good created by voluntary human labour are beyond the means (guaranteed income) of any single human and could only be given as gift, lottery or in some sort of collective agreement (team of X builds X and each gets one)

  • global industry robot sales: 2013: ~179,000

  • shift to automation slower when the capital investment of existing equipment is considerable or new equipment is considerable

    • office and knowledge work is then ripe for exploitation whereas transportation may not be