Skip to main content

Effects of automation on employment - including post-COVID-19

Primary supervisor

David Dowe

Co-supervisors

  • Dr Michelle Rendall

 Automation has affected employment at least as far back as Gutenberg, the introduction of the printing press and the effect on scribes and others. Such changes have occurred in the centuries since. In more recent times, we see electronic intelligence showing increasingly rapid advances, with examples including (e.g.) easily accessible, free, rapid and often somewhat reliable language translation. More recent advances include the increasing emergence of driverless cars. This is an area in which rapid changes continue to occur, most recently as the world both deals with and looks to emerge from COVID-19 coronavirus and we seek a sustainable path forward.

  We collect a variety of human opinions to help model and predict the various changes in various parts of the workforce. This will assist in planning, especially as the world looks to re-emerge after - and during - COVID-19.  We follow through on earlier work by Frey and Osborne (2013, 2017) and others in endeavouring to model how automation will affect the future of work.  We extend this now to endeavour to determine with the new issues of how COVID-19 coronavirus will affect the future of work and how societies, economies and workforces will respond to issues of climate change and the likely regular occurrence of bushfires such as those of the Australian 2019-2020 summer.

  We will survey a variety of people from various backgrounds in order to find their views on matters such as these - which jobs will continue as they are, be partly automated and be totally automated. We consider approximately 683 individual occupations, covering a majority of the U.S. workforce. We then consider various attributes of these occupations, as given by the Occupational Information Network (O*NET) data-base. Using a subset of these occupations, we survey a group of experts to predict (probabilistically) whether these occupations will be automated, augmented or unaffected by emerging technologies. Using this data, a classification algorithm is then trained to predict the probability that an occupation in the data-set is automated, augmented or simply unchanged at some in the future. Using the O*NET attributes and the opinions of various surveyed humans, we then endeavour to model and predict which jobs and occupations will be partly augmented or totally automated, and also - for those occupations that are predicted to be augmented or totally automated - which technologies will affect them.

It is our anticipation that the work will commence with, in parallel, the survey for collecting the data and a comparison of machine learning methods on artificial pseudo-randomly generated data. The world will see all too many changes during and in the aftermath of COVID-19. We hope this project - with its multidisciplinary team - to be one of the early projects anticipating where job markets might head.

 

References:

  Fitzgibbon, L.J., D. L. Dowe and F. Vahid (2004). Minimum Message Length Autoregressive Model Order Selection. In M. Palanaswami, C. Chandra Sekhar, G. Kumar Venayagamoorthy, S. Mohan and M. K. Ghantasala (eds.), International Conference on Intelligent Sensing and Information Processing (ICISIP), Chennai, India, 4-7 January 2004 (ISBN: 0-7803-8243-9, IEEE Catalogue Number: 04EX783), pp439-444

  Frey and Osborne (2013)

  Frey and Osborne (2017)

  P. J. Tan and D. L. Dowe (2003). MML Inference of Decision Graphs with Multi-Way Joins and Dynamic Attributes, Proc. 16th Australian Joint Conference on Artificial Intelligence (AI'03), Perth, Australia, 3-5 Dec. 2003, Published in Lecture Notes in Artificial Intelligence (LNAI) 2903, Springer-Verlag, pp269-281

  Wallace, C.S. (2005), ``Statistical and Inductive Inference by Minimum Message Length'', Springer  (Link to the preface [and p vi, also here])

  Wallace, C.S. and D.L. Dowe (1994b), Intrinsic classification by MML - the Snob program. Proc. 7th Australian Joint Conf. on Artificial Intelligence, UNE, Armidale, Australia, November 1994, pp37-44

  Wallace, C.S. and D.L. Dowe (1999a). Minimum Message Length and Kolmogorov Complexity, Computer Journal (special issue on Kolmogorov complexity), Vol. 42, No. 4, pp270-283

  Wallace, C.S. and D.L. Dowe (2000). MML clustering of multi-state, Poisson, von Mises circular and Gaussian distributions, Statistics and Computing, Vol. 10, No. 1, Jan. 2000, pp73-83.

 

Required knowledge

Mathematics at least to first-year undergraduate level, preferably more.

An ability to program.

Statistics and/or machine learning and/or data science at least to undergraduate level.

A willing to learn the necessary amount of Economics and at least an interest in the future of employment.


Learn more about minimum entry requirements.