Machine Learning: Changing the way we work.

Machine Learning (ML) won’t take people’s jobs but it will change them…

This opinion piece article was written as part of the Science Politics course at the Centre for Public Awareness of Science.

Very few jobs will go fully extinct over the next 10 years from machine learning. Those that do, are well overdue to replace humans in unsafe, wasteful and potentially deadly scenarios.

To understand the oncoming storm; we have to understand how machine learning can replace jobs and what parts of a job are ripe for change.

Machine learning is an area focused on two questions: How can one make computer systems that can improve themselves through experience? and what are the fundamental laws that govern all learning that includes learning in computers, humans and organisations [5].

Automation…

A job or occupation is essentially a bundle of tasks, you drive a truck, sort a spreadsheet, monitor a process, organise an event or program software, some tasks are more suited to automation whilst others aren’t[1]. So far the best technique with the most obvious economic potential in terms of automation is deep neural nets (DNN) [2].

Machine learning is a different technology than earlier types of automation like assembly lines or general software programming and it affects a very different set of tasks. Most occupations in most industries have at least some tasks that are suitable for machine learning, the next wave of automation and job engineering will affect a very different part of the labour force than previous areas [2].

Rather than talking about how ML will take our jobs we need to talk about how ML will redesign jobs. According to recent research, AI and robotics has the potential to increase productivity growth but will have mixed effects on labour, especially in the short run [3]. Much like the industrial revolution, some industries will do well whilst others will have a labour upheaval[3].

Slowing down the development and application of machine learning is impossible; the benefits are way too advantageous for economies and businesses. Increased productivity, safety, decreased deaths, improved research speed and more all spur machine learning on[4]. What’s important is mitigating the risk from disruption to the labour market to ensure a smooth transition into our machine learning augmented jobs.

There are many projects and initiatives underway in Australia: the 2018/2019 budget build allocated 28 million to build capability in, and support the responsible development of AI including the developments of cooperative research centres, resources to teach A.I. in schools and the creation of an AI Technology roadmap that explores the impact of AI on industries and workforce [6].

So what can we do?

It is inevitable that people will be displaced from employment disruptions and that they will happen[4]. What we can do, is focus on ensuring there are safety nets in place for when people are made redundant and ensure that the right support both financially and educationally is in place.

We have initiatives to investigate the various impacts of machine learning but not much in the realm of transition planning or safety nets; systems such as Universal Basic Income, guaranteed employment, increased Centrelink study payments, and long-term transition training are all steps on a pathway to a smoother transition, there is no single change or catch-all solution.

We must make sure that systems are in place before it’s too late so we are prepared for the redesigning and restructuring of existing jobs and ready for the emerging new jobs that require skilled and educated people whilst being capable of addressing redundancy issues.

Wondering if robots will take your job?

https://willrobotstakemyjob.com

Wondering how machines learn?

https://www.youtube.com/watch?v=R9OHn5ZF4Uo

Reference

1. Autor D, Levy F and Murnane R (2001) The Skill Content of Recent Technological Change: An Empirical Exploration. w8337, June. Cambridge, MA: National Bureau of Economic Research. DOI: 10.3386/w8337.
2. Brynjolfsson E, Mitchell T and Rock D (2018) What Can Machines Learn, and What Does It Mean for Occupations and the Economy?. AEA Papers and Proceedings, 108, 43–47
3. Furman J and Seamans R (2019) AI and the Economy. Innovation Policy and the Economy, 19, 161–191.
4. Greatorex A (2019) Machine Learning: Goodbye Society as We Know It. Towards Data Science, Available from: https://towardsdatascience.com/machine-learning-goodbye-society-as-we-know-it-c83b11542c8c (accessed 27 September 2019).
5. Jordan M and Mitchell T (2015) Machine learning: Trends, perspectives, and prospects. Science, 349(6245), 255–260.
6. OECD (2019), Artificial Intelligence in Society, OECD Publishing, Paris, https://doi.org/10.1787/eedfee77-en.