Home /

How jobs will change

Will technology destroy jobs?

Key points:

Over the past five years, a number of reports have predicted that automation could lead to large-scale destruction of jobs, with estimates reaching as high as more than 40 per cent of Australian jobs over the next decade.

No one can predict the future with certainty, but these estimates appear overblown. They seem to overstate possible job losses and do not capture the full range of impacts of technology:

The greatest impact of technology will be task change within existing jobs. Every single job can expect some change in its tasks over the next decade. Research by AlphaBeta suggests that, on average, 9.3 per cent of tasks changed in occupations over the last five years. Task change is an essential part of adapting to change. Our research suggests jobs that experience more task change have less incidence of job loss.

A small number of jobs will be entirely substituted. The most credible estimates suggest 5 to 10 per cent of jobs could be entirely substituted by technology over the next decade. The loss of any job has a personal, emotional and financial cost for the individual involved. Even if future job losses are much smaller than estimated, there is still an overwhelming imperative to assist those individuals to get back into work.

Technology will also create jobs. Technology can create jobs directly by generating the creation of new roles needed to develop or maintain the technology, or indirectly by enabling productivity improvements that enable jobs growth in other areas. However, the rate of change will vary across different sectors, and it is possible that some change may be more severe in some sectors and occupations that have been relatively immune in the past.

Australia has been adapting to technology change relatively well for decades. Rather than stoking fear about the impact of technology, we should:

support all workers to update their skills in the face of task change; and
provide more intensive support to individuals who may lose their job.

Discussion:

More Australians today have a job than at any other time in history (12.6 million working Australians).

The current wave of emerging technologies (like artificial intelligence, robotics and automation[1]) produces a number of new policy challenges, including the impact on Australian workers.

Technological change has been occurring in Australia for decades. Many technologies have taken over tasks that were previously done by people: from the loom and the harvester, to mass textile production and automatic supermarket checkouts.

The introduction and adoption of information and communication technologies over the last three decades is a good example. The computer and the internet have fundamentally changed the nature of work, jobs and workplaces. However, following the widespread adoption of information and communication technologies, there was no discernible negative impact on employment across G20 countries.^[2] Dynamic labour markets help organisations and individuals to gradually adjust, and incrementally move the allocation of people and resources to more productive uses.

The cumulative weight of evidence demonstrates that, when labour-saving technology is adopted, labour is re-directed to other areas of demand. In this way, Australia’s labour markets are reasonably effective at accommodating change.

Some commentators suggest that improvements in technology will mean swathes of work can be delegated entirely to technology, leading to a massive surplus of labour and potential re-imagining of how people will allocate their time to work or leisure.^[3] But, based on historical evidence, large surpluses of workers seem unlikely. In spite of the transformative impact of information and communications technologies, hours worked per capita in Australia has stayed fairly stable (outside of economic downturns) since the 1960s.^[4] Treasury projects that hours worked per person will remain essentially stable over the coming decades and, if anything, an ageing population is more likely to reduce the proportion of traditional working age Australians and the participation rate.^[5]

Although the overall, aggregate amount of work is unlikely to change, selected cohorts of workers may be disadvantaged by developments in technology. To understand the full impact of technological change, we need to consider:

jobs that will remain but will change;
jobs that will be entirely substituted by technology; and
new jobs that will be created by technology.

Technology will result in changes in tasks

Every single job in Australia will likely experience some change in tasks over the next decade.

This is not a new phenomenon. Almost every working Australian gradually adapts to change in their job as a matter of course. It is very difficult to imagine any job in Australia that does not involve a computer, a smartphone, the internet or some form of software at some stage.

Analysis by AlphaBeta suggests that, between 2011 and 2016, the average level of task change within occupations was around 9.3 per cent (Figure 1). In other words, Australian workers now spend about half a day a week doing tasks that people in the same job were not doing five years ago.

Here are some examples of the task changes that have already occurred:

Accountants are spending less time computing data and more time resolving clients’ problems.
Registered nurses are spending less time recording patient histories, and more time monitoring patients.
Product assemblers are spending less time assembling machinery and more time reviewing and learning.

Many of the tasks that shift from being performed by people to technology will be risky, routine or repetitive. Workers currently performing these tasks have the lowest reported levels of job satisfaction.^[6]

The remaining tasks that will experience greater demand are likely to be tasks that are not well performed by technology, such as solving problems, generating creative ideas and uniquely human interactions. Tasks less likely to be automated and therefore may experience higher demand include:

perception and manipulation tasks (such as identifying objects and moving them in an unstructured or cluttered working space, or personal services like beauty therapy, or fine motor functions that require detail [like the work of an electrician])
creative intelligence tasks (such as artistic design, musical composition and cheffing)
social intelligence tasks (such as negotiation, persuasion and care), and
problem solving tasks.^[7]

Considering the significant impact technology could have on every single job, an ability to adapt to changing tasks will be essential for every working Australian.

There are many benefits to Australians from adapting to task change: AlphaBeta’s research finds that jobs that experience more task change have less incidence of job losses.

But, we also need to be alert to jobs where the task change may be so great that existing workers are at risk of involuntary transition (job loss).

The education and training system has an important role to play (especially when tasks are changing more rapidly), but the incremental change in tasks also highlights the critical importance of on-the-job learning.

A small number of jobs will be entirely substituted by technology

In a small number of instances, it is possible that technology will substitute sufficient tasks to make an entire job obsolete. This has previously occurred with jobs like typists or photo developers.

If the adjustment occurs quickly, some individuals could find themselves out of work in the short term.

The loss of any job has a personal, emotional and financial cost for the individual involved.

Over the long term, jobs that are routine (whether cognitive or manual) are particularly at risk.^[8] In fact, they already have been declining for some time:

In the last 30 years, routine manual jobs (eg. machinery operators) have fallen from 40 to 30 per cent of the labour market, while the proportion of routine cognitive jobs (eg. clerical workers) has fallen from 27 to 23 per cent.
At the same time, the number of non-routine manual jobs (eg. hospitality workers) has risen from 6 to 11 per cent, as have non-routine cognitive jobs (eg. professional occupations) rising from 27 to 36 per cent.^[9]

There have been many high-profile estimates of gross job losses, that range from the very small to the very large. A summary of the estimates is provided in Figure 2.

^[10]^[11]^[12]^[13]^[14]^[15]^[16]

Some of these estimates overstate the potential job losses. For example, one critique suggests the methodology in Frey and Osborne (2013) – among other things – assumes that, if one task in a job can be automated, the entire job will be substituted.[17] We believe it is much more credible to expect that, if one task within a job is automated, the job will likely remain and the worker’s time will be devoted to a different, more productive task.

For this reason, the most credible estimates seem to suggest the job displacement will be at the smaller end (between 5 and 10 per cent over the next decade).

Even if future job losses are much smaller than estimated, this does not diminish the need for a proper discussion on what we need to do to prepare. There is still an overwhelming imperative to assist those individuals to get back into work. The personal, emotional and financial cost for the individual can be so significant that, even if the aggregate numbers are smaller than expected, there is still more that can be done to assist those individuals.

The timing and location of job substitution can be difficult to predict. The speed at which labour-saving technology is adopted will depend on a range of factors (including cost, social acceptance and regulatory requirements) – not just the availability of the technology. The rate of change will differ across different sectors, and it is possible that some change may be more severe in some sectors and occupations that have been relatively immune in the past.

Technology will also create jobs

We can also expect that technology will directly and indirectly create jobs.

Some of the jobs that would be directly created would likely include those that:

develop, maintain and propagate the technology. For example, greater use of artificial intelligence will necessitate jobs for people who are ‘trainers’ (who teach artificial intelligence systems how to act), ‘explainers’ (who translate artificial intelligence results into language for people to understand) and ‘sustainers’ (who maintain artificial intelligence systems and ensure ethical use)^[18], and
use or complement the technology: for example, in Australia, there has been a rise in demand for photographers at the same time there has been a decline in photo developers and printers^[19]

Technology also indirectly creates jobs, primarily through improving productivity. Productivity is important because it is the key determinant of Australia’s long-run living standards. Increasing productivity is critical to increasing real incomes over time.

As technology improves productivity, the resulting higher incomes create new or increased demand for goods and services generally, and free up resources to be used for more productive purposes.

The displacement effect from technology in the past has been offset by improvements in productivity.^[20]

There are many estimates about the potentially large productivity benefits to be gained from deployment of technology. For example, McKinsey embracing digital technologies could add between $140 billion and $250 billion to the Australian economy by 2025.[21]

The direct and indirect creation of jobs by technology is likely to lead to net jobs growth, even as jobs may be displaced by technology over the long term.

Glossary

Artificial intelligence
Technologies that analyse data and language in new ways to recognise complex patterns and improve their own analysis.

Automation
Automation could extend to any technology that allows a task to be completed without a human performing it. Robotic process automation is a type of software that can complete a task without a human performing it.

Internet of Things
The connection of physical objects to the Internet (potentially through sensors or other technologies), allowing objects to communicate data, and be controlled remotely.

Robotics
The development of physical machines that can replicate human actions.

Will all new jobs be high-tech, highly skilled jobs?

Key points:

Much of the public discussion of work in the future focusses on the new jobs that will be created by technologies like artificial intelligence, robotics, cyber security and data. While these types of jobs hold exciting potential, they are likely to represent a relatively small proportion of the Australian labour market.
Jobs growth is likely to occur in many industries and parts of the labour market, but the health, aged and child care sectors are projected to experience the most growth over the next five years.
Technology will require a higher level of skill and knowledge generally across the labour market. However, in the short term, Australia is expected to experience jobs growth at all skill levels.

The skills needs of the Australian labour market have progressively been increasing for some time.[22]

The average increasing level of skill has been driven by multiple factors:

Higher levels of skill have been required in sectors to operate and manipulate new technology. For example, in mining, as processes become increasingly automated.

Advances in technology are biased towards skills. Over the 20^th century, technological development raised demand for skilled workers, resulting in more jobs and wage premiums.^[23]
Shifts in the industry composition of employment has led to stronger growth in higher-skilled jobs in industries like professional services and lowest growth in lower-skilled jobs, such as machinery operators, labourers, and clerical workers.

The services sectors dominate the Australian labour market and this is projected to continue. The number of workers in these sectors is already a sizeable proportion of the labour force and the Department of Jobs and Small Business projects these sectors will continue to grow at a faster rate than other sectors. Over the next five years, the fastest growing sectors are projected to be health, aged care and child care (14.9 per cent), education (11.2 per cent) and professional, scientific and technical services (10.2 per cent) (refer Figure 1).

The projected growth in the health care and social assistance sector alone (an additional 250,000 people over five years) outnumbers the current size of some industries.

As community demand for services such as health care continues to rise and generate employment in that sector, it will also be important to ensure we are growing jobs in other parts of the economy that are less reliant on government funding.

When we examine the projections by skill level, it is clear that there will continue to be strong growth at all skill levels – but especially the highest skill level (refer Figure 2).

Note: What are Skills Levels 1 to 5?

The Australian Bureau of Statistics has assigned every job in Australia a ‘Skill Level’, the level of skills required to perform that job. Skill Level 1 is the highest, Skill Level 5 is the lowest.
The majority of occupations at Skill Level 1 are in the managers and professionals occupational groups. Examples of specific jobs include: economists, dentists, production managers and livestock farmers.
The majority of occupations at Skill Level 2 are spread across the managers, technicians and trades workers, community and personal service worker occupational groups. Examples of specific jobs include: retail managers, safety inspectors and ICT support technicians and paramedics.
The majority of occupations at Skill Level 3 are in the technicians and trades workers occupational groups. Examples of specific jobs include: florists, cooks and plumbers.
The majority of occupations at Skill Level 4 are in the sales workers, machinery operators and drivers, and labourers occupational groups. Examples of specific jobs include: visual merchandisers, delivery drivers and product quality controllers.
A small subset of roles can also be found in the community and personal service workers and clerical and administrative workers occupational groups. Examples of specific jobs include: dental assistants and bank workers.
The majority of occupations at Skill Level 5 are in the sales workers, machinery operators and drivers, and labourers occupational groups. For example, sales assistants, concreters and packers are skill level five occupations.

Every one of the top 20 highest employing occupations are projected to need more workers in five years’ time than they employ today, even though the rates of growth will vary (refer Figure 3).

The Department of Jobs and Small Business projects the greatest growth in occupations between 2018 and 2023 will be in the following minor occupational groups:

aged and disabled care workers (net job growth of 69,200)
registered nurses (51,400)
child carers (27,600)
software and application programmers (25,500), and
waiters (21,800).

Sales and hospitality occupations employ many people, including a large share of new entrants to the labour market. There are also large ‘outflows’ in workers from these industries, suggesting that many young workers begin their working lives in these industries before moving to other occupations.^[24] The strong growth in sales assistants and hospitality workers suggests there will continue to be entry-level opportunities in the labour market.

How well have we dealt with labour market change in the past?

Australia has dealt well with labour market change in the past. Ultimately it has been the economic cycle (recessions and slowdowns), not structural change that has had the largest observable impact on the labour market.

This underscores the need to keep the economy strong.

The Australian labour market has seen significant change over the past 30 years. These changes include a large rise in the share of Australians in the labour force; and also shifts in the share of employment across industries and sectors of the economy.

As well as these structural changes, the economic cycle has also had a big impact on the labour market – most notably the recession of the early 1990s and also the slowdown during the global financial crisis (GFC). It has ultimately been these slowdowns and recessions that have had the largest observable impact on the overall health of the labour market (Figure 1).

The weakness in the jobs market during these slowdowns has been dramatic, and vast numbers of Australians have been adversely affected.

For example, in the early 1990s recession the unemployment rate increased from 5.8 per cent to 11.2 per cent over three years. In today’s labour market, an unemployment rate of 11.2 per cent would correspond to 1.5 million people out of work.
It then took over a decade to return the unemployment rate to the low seen just prior to the 1990s recession.
During the GFC, the unemployment rate increased from 4 per cent to 5.9 per cent in less than a year, with the number of unemployed Australians increasing by 219,000.[25]

The personal toll of labour market weakness during recessions and economy slowdowns underscores the importance of a strong economy. It is only through a strong economy that we can, ultimately, keep the unemployment rate low and Australians in jobs.

Conversely, the general robustness of the Australian economy over the past 27 years is reflected in our current unemployment rate of 5 per cent – a low rate by long-run standards.

There have also been significant changes in the shares of employment across broad sectors of the economy over the past 30 years. The share of Australian employment in both market and non-market services has risen since the mid-1980s. Being a services economy is not a new development for Australia, with the services sector (market and non-market) comprising a majority of employment in the mid-1980s.

There has, however, also been a significant fall in the share of employment in the goods production sector.

While the absolute number of those employed in the goods production sector has not fallen by much (there are 160,000 fewer people employed in this sector than 30 years ago), the reduction in the share of employment in the sector is stark.

In fact, if the share of employment in the goods production sector had remained the same over the past 30 years then there would be 1.6 million more people employed in that sector today (and fewer people employed in other sectors).

The ability of the Australian economy to accommodate such a shift in employment between sectors – in effect, re-distributing over one and a half million jobs in the goods production sector for jobs in services – suggests that we have been able to deal with labour market change in the past.

Of course, for the individual workers involved, the adjustment described above is likely to have been quite dislocating. The experience of the economy in aggregate is unlikely to reflect the experience of every individual within it. As a result, there is a role for government to play in smoothing the adjustment process for individual workers through ensuring, for example, access to education and training.

It is also important to note that the period we have considered (post 1985-86) generally corresponded with an opening up of the Australian economy and (until recently) increases in labour market flexibility.

A less flexible labour market is likely to have been less successful in accommodating such a degree of change.