The spread of advanced automation, robotics, and artificial intelligence is transforming the way manufacturers work. Manufacturers’ Monthly explores how these development are impacting the workforce.
In the manufacturing sector, technological change and transformations in production processes are pointing towards changes in business models. In particular, developments in advanced automation, robotics artificial intelligence (AI), machine learning, sensor technology, data analytics, and 3D printing are altering the way in which operations are carried out. That there will be impacts upon the makeup of the workforce is obvious; but the precise nature of the long-term changes to work in the manufacturing sector is debated.
The factors leading to the intensification of industrial automation in recent years, including the growth of AI, are expansions in the capacity for data storage and advances in data analytics, computational power and internet connectivity. Along with these, improvements in transferring digital instructions to the physical world, such as robotics and 3D printing are, according to some experts, on the precipice of transforming the world of work.
According a recent report by Oxford Economics, “How Robots Change the World”, while the ever- faster adoption of robots will lead to greater productivity and economic growth, and create jobs in hitherto non-existent roles and industries, business models across a variety of sectors will be disrupted, with tens of millions of jobs lost as human workers are displaced by machines.
“The number of robots in use worldwide multiplied three-fold over the past two decades, to 2.25 million,” said Oxford Economics’ CEO and chief economist Adrian Cooper. “Trends suggest the global stock of robots will multiply even faster in the next 20 years, reaching as many as 20 million by 2030, with 14 million in China alone. The implications are immense, and the emerging challenges for governments and policymakers are equally daunting their scale.”
In the manufacturing sector alone, the report states, worldwide job losses are set to reach 20 million.
This figure is 8.5 per cent of the global manufacturing workforce. The jobs most at risk are those in low-skill areas and those with lower productivity. According to economic modelling by Oxford Economics, recent improvements in New South Wales’ manufacturing productivity mean that the impact of future robot densification in the state will be more or less muted. South Australia, on the other hand, is much more vulnerable. The state is Australia’s most intensive in terms of manufacturing, but it also has low levels of manufacturing productivity.
Manufacturing is already an intensely automated sector, currently accounting for around 86 per cent of the total global operational stock of robots according to the International Federation of Robotics. The Oxford Economics report suggests that there are three main drivers behind the adoption of robotics in manufacturing: price, innovative applications, and consumer demand. The continued exponential growth in the processing power
of microchips and the increased capacity of networks mean that the per-unit cost of many technological components has been dropping. And, with developments in AI, robots and machines are making more informed decisions from the data that is, in ever greater volumes, gathered and stored across industrial sectors and systems. It is these developments, too, that are helping drive the adoption of robotics and automation across a broader swathe of manufacturing sectors.
In recent years, collaborative robots, or cobots, enabled by AI and cloud technologies, have begun to emerge, bridging the divide between manual assembly performed by skilled human workers and automated production. This has the potential to deeply transform the way work is performed in the manufacturing sector. Speaking earlier this year at AUSPACK 2019 in Melbourne, Universal Robots’ Peter Hern made the case that collaborative robots were, in fact, enabling production to be focussed back towards the human workforce. “When we talk about robots, a lot of people ask, ‘Are we losing our jobs to robots?’ Absolutely not. In fact, what we find is that the human robot collaboration is actually 85 per cent more productive that humans or robots alone,” he said. “It enables companies to reposition their staff, their workers, from dull, dangerous, and mundane roles into more productive roles and higher value roles.”
The Oxford Economics report itself states that it would be “simplistic” to characterise robotisation and the adoption of advanced automation as merely a “destroyer” of jobs. It will instead lead to a “robotics dividend”, characterised by lower prices for manufactures, higher real incomes, and higher tax revenues. A one per cent increase in the stock of robots per worker in the manufacturing sector leads to a 0.1 per cent growth in output per worker.
“This confirms our hypothesis,” the report’s authors state. “[By] displacing automatable jobs in manufacturing, robots free up many workers to contribute productively elsewhere in the economy, as they meet the demands generated by lower prices for manufactured goods.”
A July talk at Deloitte Australia in Sydney on the future of the workforce saw the consulting company’s chief data scientist Gavin Whyte attempting to calm fears, propelled by the media, that the adoption of AI would lead to the disappearance of jobs. “Yes, AI will impact business, permeate industry and affect your roles, but in a good manner,” Whyte said.
According to Whyte, as AI continues to permeate the economy, it will transform the way work is carried out – not by displacing the workforce, but by freeing workers from mundane tasks and activities to carry out higher-order tasks.
Drawing attention to the way the internet has transformed economic activity since 1995, Whyte said that as the adoption of computers and the internet intensified, the cost of search and of digital transfers dropped dramatically. Similarly, as the processing power of computers expands, the cost of what Whyte calls “arithmetic” – algorithmic processing – drops. “And what happens when the cost of an item drops? We use more of it. The cost of search has dropped drastically, so we use more of it – we use Google search for so many of our activities today – and the cost of arithmetic has dropped tremendously, so we use more of it in areas like demand forecasting, and in using computers to record and transfer ever-larger volumes of data.”
Moving more directly on to the topic of AI, Whyte said that the AI the cost of prediction had dropped dramatically in recent years. “It has plummeted to a point where prediction is now cheaper, quicker, faster. And, when something becomes cheaper – like the internet, like arithmetic – we start using more of it for a whole range of purposes,” he said. Prediction is essentially the use of information to generate information that you don’t currently have. If you start applying AI to prediction problems, you get a better outcome with the resulting information, and we can start utilising the power of prediction in areas that have, historically, not been problems of prediction at all.”
And, according to Whyte, that means AI is making our jobs, and our lives, more convenient. “We require convenience. And convenience comes at a cost.” If we use AI to make life and work more convenient, said Whyte, it will impact some roles; but as with every other industrial revolution, where more advanced techniques boosted productivity and opened up new roles in production, the development of AI will open up the potential for new roles along with the means of helping and improving workers in their existing roles.
“The idea pushed in the media today is that ‘AI is going to come and automate your roles and thousands are going to lose their jobs’. But if you really look into the heart of the matter, this is not the case,” he said.
According to Whyte, as the cost of machine prediction drops, enabling more and greater volumes of data to be processed, it also enables better further, better human judgement. “A machine makes predictions. But who is going to use those predictions? Humans. We need to make judgements from them,” he said. Whyte argued that the speed, quality, volume, soundness and accuracy of human judgement would increase as machine prediction is applied to more spheres of work, making AI a complement to human workplace roles. “AI is not going to replace you – AI is going to improve your judgement and improve outcomes.”
Getting on board with the AI revolution, however, will require training that prepares the workforce for jobs of the future. In September last year, the Australian Industry (Ai) Group released a report on its survey of 300 businesses, 75 per cent of which reported skills shortages. Among the most serious gaps were in skills utilising data, automation and AI.
Federal industry, science and technology minister Karen Andrews recently attended G20 meetings in Japan regarding the spread and adoption of AI. Speaking about the meetings, Andrews said that AI was increasingly seen as the “petroleum of the twenty-first century”, with all member countries regarding AI as a critical issue for the future.
“I think it’s fair to say that the impact of AI will make a significant difference to the way we work and the way we live, and it’s important that we make sure that we get the policy settings right,” Andrews said.
The adoption and development of machine learning and AI by industry is being encouraged via the federal government’s Taking Local Businesses Global initiative, $29.9 million will support Australian businesses and workers to develop artificial intelligence and machine learning capabilities. This includes co-funding for Cooperative Research Centre grants for industry-led research.
Andrews said that is important that workers’ fears and concerns that AI adoption will lead to a reduction in job opportunities, or even make them redundant, be taken into account
in the development of policy, which should focus on training people to help them transition to other jobs.
In a submission to the federal government’s Senate Select Committee on the Future of Work and Workers last year, University of Adelaide researchers indicated that over the period of increasing automation, unemployment has not been correlated with the increase, while worker productivity, their corresponding income and the broader gross domestic product has risen sharply over the same period.
“While automation reduces the need for some occupations, it creates the need for new ones, which are often higher paid,” the submission states. “The car industry uses the largest number of industrial robots in the world. But there was little correlation between automation and the change in labour share in the car industry between 2004 and 2014.”
While AI has the potential to increase Australia’s productivity, the submission indicated that, currently, two thirds of Australian organisations are having difficulties in finding suitable staff to lead AI technology integration and 75 per cent of IT decision makers feel that the executive team in their organisation needs formal training on the implications of AI technologies.
The researchers said that Australia is peculiarly well placed to capitalise on the current situation, but, also, that it needs to do so quickly, as other countries are investing heavily in these areas.
“As AI-driven automation lowers the cost of production, Australia could once again become competitive in manufacturing goods that are currently produced cheaply elsewhere because of low wages in other countries,” the submission states. “But we will need to encourage investment in new-generation automation to take advantage of this new trade opportunity, and ensure we have the education, training and research in place to capitalise.”
CSIRO’s data innovation group Data61 is currently leading the development of an AI roadmap and ethics framework on behalf of the federal government.
Chief executive of CSIRO, Dr Larry Marshall, in a presentation last year at the AFR Innovation Summit, said that AI was going to transform the future of work, and that staying ahead in AI investment, research and training would ensure Australia stays at the forefront of job creation in new industries, predicting and preparing for them before they emerge.
“To ensure a long and prosperous future for Australian industries, we need to ensure Australia grows its pipeline of AI-literacy – which is really STEM literacy,” Marshall said.
“I’ve heard people say that coding is all kids will need to know in the future, but I disagree. What is most important is to instil our children with the ability to learn how to learn. That is the beauty of science. It teaches you how to tackle problems and gives you a toolkit to approach any problem.
“The most important thing to remember, in the face of all of the hype, is that for machine learning technology to work well, it needs to ‘learn’, and it is our role to help give this technology real intelligence.”