Economists and policy makers worry that the rapid advance of artificial intelligence (AI) and automation technologies could seriously disrupt labour markets. While technology might boost the productivity of some workers, it may well replace others entirely, and will probably change all occupations at least to some degree. How can we benefit from these technologies, while also avoiding painful social disruption and unemployment, or amplifying already troubling economic inequality? We should anticipate a range of complex effects, as the consequences of new technology cascade through society on many levels, altering occupational skill requirements and career mobility, or changing workers’ social identities.
As LML External Fellow Hyejin Youn and colleagues note in a paper, researchers and policy makers currently lack the necessary information to forecast such effects. To give researchers a better chance, the authors suggest we urgently need to collect more detailed data to give a more accurate picture of real-time changes in the labour market. Useful sources of data include unstructured skills data from resumes or job postings, for example, as well as others indicators (e.g. patent data) reflecting recent technological change. Improved data collection, they suggest, would enable the use of a variety of data-driven tools, including machine learning applications, to more accurately model the complexity of labour systems. Only with better data will we have a chance to chart a path into a more automated future with some awareness of the problems and opportunities likely to arise.