by Frey & Osborne on the future of employment In 2013 the economist Carl Frey and the ML coder Michael Osborne, both at Oxford, published the working paper, ‘The future of employment: How susceptible are jobs to computerisation?’. The headline finding of the paper was that in the near future 47 per cent of total US employment was at high risk of displacement by AI and robotics, and 33 per cent at low risk … In 2015, responding to Frey and Osborne, the Bank of England applied the same approach to the UK economy and produced equivalent figures … The headline findings of both Frey and Osborne and the Bank of England are troubling (47 per cent and 35 per cent of total employment at high risk of displacement). They invite anxiety. But what has actually been
Topics:
Lars Pålsson Syll considers the following as important: Economics
This could be interesting, too:
Lars Pålsson Syll writes The total incompetence of people in charge of the US economy
Lars Pålsson Syll writes Ergodicity — a questionable assumption (wonkish)
Lars Pålsson Syll writes Vägval i finanspolitiken
Lars Pålsson Syll writes Economics — a dismal and harmful science
Frey & Osborne on the future of employment
In 2013 the economist Carl Frey and the ML coder Michael Osborne, both at Oxford, published the working paper, ‘The future of employment: How susceptible are jobs to computerisation?’. The headline finding of the paper was that in the near future 47 per cent of total US employment was at high risk of displacement by AI and robotics, and 33 per cent at low risk … In 2015, responding to Frey and Osborne, the Bank of England applied the same approach to the UK economy and produced equivalent figures … The headline findings of both Frey and Osborne and the Bank of England are troubling (47 per cent and 35 per cent of total employment at high risk of displacement). They invite anxiety. But what has actually been assumed and achieved? …
Arguably, the model and the findings provide a baseline for discussion. However, consider again how the numbers are produced: a claim is made that a specific percentage of all occupations in a database are in the near future at high ‘risk’ of displacement by technology, but simultaneously we are informed
that the figures take no account of work modification, new jobs created, and the
actual socio-economic environment for displacement within which the developing technologies will be substantively influenced and taken up. The method meanwhile is internally related to refinements of tech-expert decisions on classifications. The assumptions, therefore, not only lack realism, the numbers can have no real-world analogue now or in the future. Even if the future levels of (un)employment for occupations at some point in time coincide, all of what is put aside in constructing the model will have been influential in producing that outcome and so it would be more reasonable to describe ‘coincide’ as coincidence in the ordinary language sense. Future reality will not be confirming the findings based on the method and there is something dubious about the typical way of referring to the kind of repeated running of simulations that is inherent to the approach as ‘experiment’. No causal powers are isolated and manipulated to explore or test some real relation. This is mathematics, it is computation, but is it science and is it social science?
Morgan’s must-read more than anything else reveals how numbers and ‘data’ produced using questionable assumptions and methods are far from the neutral and objective things they standardly are portrayed as when economists present their model-based contingent quantifications of a fundamentally uncertain future.