As machines increasingly replace the labor of the workplace, we will all need to adjust to get the benefits.
Automation and artificial intelligence are transforming businesses and will drive economic growth through contributions to productivity. They will also help address the social challenges of “moonshot” from health to climate change.
At the same time, these technologies will change the nature of the work and the workplace itself. The machine will be able to perform more tasks that humans can accomplish, assist humans in their work, and even perform tasks that humans cannot accomplish. As a result, some professions will decline, some other professions will thrive, and more careers will change.
Although we believe that there will be enough work to do (except in extreme cases), society needs to deal with major labor transfers and dislocations. Workers need to master new skills and adapt to increasingly powerful machines in the workplace. They may have to move from a declining career to an increasingly prosperous career, and in some cases even need to change careers.
This executive brief draws on the latest research from the McKinsey Global Institute, explores the prospects and challenges of workplace automation and artificial intelligence, and outlines some of the key issues that policymakers, companies and individuals need to address.
1. The accelerated process of artificial intelligence and automation is creating a lot of opportunities for business, the economy and society.
2. How artificial intelligence and automation will affect work
3. Transformation and challenges of the main workforce
4. Ten things to solve
The acceleration of artificial intelligence and automation is creating plenty of opportunities for business, the economy and society.
Automation and artificial intelligence are no strangers, but the latest advances in technology are driving cutting-edge work that machines can do. Studies have shown that society needs these advances that provide value to businesses, promote economic growth, and achieve unimaginable breakthroughs in some of the most difficult social challenges.
Technology process acceleration
In addition to traditional industrial automation and advanced robotics, a new generation of more powerful automation systems is also available in a variety of environments, from autonomous vehicles on the road to automatic checkout at the grocery store. Most of the advancements are driven by improvements in systems and components, including machinery, sensors, and software. Artificial intelligence has made particularly significant progress in recent years as machine learning algorithms have become more sophisticated and have taken advantage of the tremendous improvements in computing power and the exponential growth of data that can be used to train them. Amazing breakthroughs are frequently becoming headlines, many of which involve computer vision, natural language processing, and complex games (such as GO) that go beyond human capabilities.
The potential to change businesses and promote economic growth
These technologies have generated value in a variety of products and services that cross-industry companies use in a range of processes to recommend personalized products, discover anomalies in production, identify fraudulent transactions, and more. The development of the latest generation of artificial intelligence technologies, including technologies that address classification, estimation, and clustering issues, is expected to bring more significant value. Our analysis of hundreds of individual smart use cases found that the most advanced deep learning techniques for deploying artificial neural networks can earn as much as $3.5 trillion to $5.8 trillion per year, or 40% of the value created by all analytics technologies ( See chart 1).
As the current ageing and reduced birth rates are dragging on, the deployment of artificial intelligence and automation technology can make a significant contribution to improving the global economy and increasing global prosperity. As a key driver of economic growth, labor productivity growth has slowed in many economies, from 2.4% in the US and major European economies a decade ago to an average of 0.5% in 2010-2014, which was previously prosperous. The result of the 2008 financial crisis following the decline in productivity. Artificial intelligence and automation have the potential to reverse this downward trend: productivity growth can reach 2% per year over the next decade, with 60% of growth coming from digital opportunities.
Helps solve the potential of several social moonshot challenges
Artificial intelligence has also been used in everything from materials science to medical research and climate science. Applying these technologies to these areas and those areas can help solve the challenges of social moon landing programs. For example, Geisinger's researchers have developed an algorithm that can reduce the time to diagnosis of intracranial hemorrhage by up to 96%. At the same time, researchers at George Washington University are using machine learning to more accurately measure the climate model used by the Intergovernmental Panel on Climate Change.
Challenges still exist before these technologies can realize their potential for economic and social benefits
Artificial intelligence and automation are still facing challenges. These limitations are partly technical, such as requiring a large amount of training data, and it is difficult to "generalize" the algorithm across applications. Another challenge is the use of artificial intelligence technology. For example, explaining the decisions made by machine learning algorithms is technically challenging, which is especially important for use cases involving financial lending or legal applications. Potential deviations in training data and algorithms, as well as data privacy, malicious use, and security, are issues that must be addressed. Europe leads the new general data protection regulations, which provide users with more data collection and use rights
Another challenge involves the ability of organizations to adopt these technologies, where people, data availability, technology, and preparation procedures often make it difficult. The adoption of various sectors and countries has been uneven. The financial, automotive and telecommunications industries are leading the adoption of artificial intelligence. Among countries, US investment in artificial intelligence ranked first in 2016, from $15 billion to $23 billion, followed by Asian investments of $8 billion to $12 billion, and Europe's backwardness, with investments of $3 billion to $4 billion.
How artificial intelligence and automation will affect work
Even if artificial intelligence and automation bring benefits to businesses and society, we need to be prepared for major disruptions in our work.
About half of the activities (not work) performed by workers can be automated
Our analysis of more than 2,000 work activities in more than 800 occupations shows that certain categories of activities are easier to automate than others. They include physical activity in highly predictable and structured environments, as well as data collection and data processing. These account for about half of the activities people do in all sectors. The least vulnerable categories include managing others, providing expertise, and communicating with stakeholders.
Almost all occupations will be affected by automation, but only about 5% of the technologies already shown are fully automated. The part of the more professional activities is automatable: we find that about 30% of the 60% of the professions can be automated. This means that most workers – from welders to mortgage brokers to CEOs – will work with fast-growing machines. The nature of these occupations may change as a result.
Unemployment: Some occupations will see a sharp decline by 2030
Automation will replace some workers. We found that between 2016 and 2030, about 15% of the global workforce, or about 400 million workers, may be unemployed due to automation. This reflects our medium scenario in predicting the speed and range of adoption. In the fastest scenario we modeled, this figure rose to 30%, or 800 million workers. In our minimum adoption scenario, only about 10 million people are unemployed, close to zero percent of the global workforce (see Exhibit 2).
广泛的范围强调了影响人工智能和自动化采用的步伐和范围的多种因素。自动化的技术可行性只是第一个影响因素。其他因素包括部署成本; 劳动力市场动态(包括劳动力供给数量、质量和相关工资); 超出有助于采用商业案例的替代劳动力的利益; 以及最后，社会规范和接纳。由于上述因素的差异，尤其是劳动力市场动态，各国和各部门的采用将继续存在显著差异：在工资水平相对较高的发达经济体，如法国、日本和美国，到2030年自动化可能会取代20%至25 %的劳动力，在中等采用情景中，印度为两倍之多。
就业：同时，也将创造就业机会
即使工人被机器所取代，工作和工作岗位需求也会随之增加。我们根据工作需求的几个催化剂制定了到2030年的劳动力需求方案，包括收入增加、医疗保健支出增加，以及基础设施、能源和技术开发和部署方面的持续或加强投资。这些方案显示，到2030年，全球劳动力(5.55亿和8.9亿个工作岗位)的额外劳动力需求范围将增加21%至33%，远远超过失去的工作岗位数量。一些最大的收益将来自印度这样的新兴经济体，那里的劳动年龄人口正迅速增长。
包括商业活力和生产力增长在内的额外经济增长也将继续创造就业机会。如果过往经验是一个指向标，那么我们目前无法想象的许多其他新职业也将出现，并且到2030年可能占创造就业机会的10%。此外，技术本身在历史上一直扮演净就业创造者的角色。例如，在20世纪70年代和80年代引入个人计算机不仅为半导体制造商，而且为所有类型的软件和应用程序开发人员，客户服务代表和信息分析师创造了数百万个工作岗位。
工作改变：随着机器辅助工作场所的人力劳动，比失去或获得的工作岗位的更多的工作岗位将发生改变
随着机器辅助人力劳动，部分自动化将变得更加普遍。例如，能够以高精度读取诊断扫描的人工智能算法将帮助医生诊断患者病例并确定合适的治疗方案。在其他领域，具有重复性任务的工作可能会转向管理和排除自动化系统故障的模型。在零售商亚马逊，以往升调和堆放物品的员工正在换成机器人操作员，用以监控自动臂并解决诸如物体流通中断等问题。
主要劳动力的转型和挑战
虽然基于我们大多数的方案，我们预计在2030年将有足够的工作来确保充分就业，但伴随采用自动化和人工智能的转变将是十分重要的。 职业组合将发生变化，技能和教育要求也将发生变化。 需要重新设计工作，以确保人类能最有效地与机器一起工作。
工人将需要不同的技能才能在未来的工作场所中茁壮成长
自动化将加速过去15年中我们所见证的所需劳动力技能的转变。 对程序设计这样的先进技术技能的需求将迅速增长。 社交、情感和高级认知技能，如创造力、批判性思维和复杂的信息处理，也将产生不断增长的需求。基本的数字技能需求一直在增加，这种趋势将继续并加速。 对物理和手工技能的需求将下降，但在许多国家，2030年仍物理机手工技能仍将是最大的劳动力技能类别(见图表3)。这将对已经存在的劳动力技能挑战以及对新的资格认证系统的需求造成额外压力。 虽然一些创新的解决方案正在兴起，但仍需要能够与挑战规模相匹配的解决方案。
The broad scope highlights a number of factors that influence the pace and scope of artificial intelligence and automation adoption. The technical feasibility of automation is only the first factor. Other factors include deployment costs; labor market dynamics (including labor supply quantity, quality, and related wages); benefits beyond the alternative labor force that helps to adopt business cases; and, finally, social norms and acceptance. Due to the above factors, especially the labor market dynamics, the adoption of countries and departments will continue to be significantly different: in advanced economies with relatively high wages, such as France, Japan and the United States, automation may replace 20 by 2030. From 2% to 25% of the workforce, India is twice as much in the medium-use scenario.
Employment: At the same time, it will also create jobs
Even if workers are replaced by machines, the demand for jobs and jobs will increase. We have developed a workforce demand plan for 2030 based on several catalysts for job requirements, including increased income, increased health care spending, and sustained or enhanced investment in infrastructure, energy and technology development and deployment. These programs show that by 2030, the global labor force (55 million and 890 million jobs) will increase the range of additional labor demand by 21% to 33%, far exceeding the number of lost jobs. Some of the biggest gains will come from emerging economies like India, where the working-age population is growing rapidly.
Additional economic growth, including business viability and productivity growth, will continue to create jobs. If past experience is a target, many other new careers that we can't imagine now will also emerge, and by 2030 may account for 10% of job creation. In addition, technology itself has historically played the role of a net job creator. For example, the introduction of personal computers in the 1970s and 1980s created millions of jobs not only for semiconductor manufacturers, but also for all types of software and application developers, customer service representatives, and information analysts.
Work change: With the labor of the machine-assisted workplace, more jobs than the lost or acquired jobs will change.
With machine-assisted labor, some automation will become more common. For example, an artificial intelligence algorithm capable of reading a diagnostic scan with high precision will help a doctor diagnose a patient's case and determine an appropriate treatment plan. In other areas, work with repetitive tasks may shift to models that manage and troubleshoot automation systems. At retailer Amazon, employees who have upgraded and stacked items in the past are being replaced by robot operators to monitor the robotic arm and solve problems such as interruptions in the flow of objects.
Major labor force transformation and challenges
While based on most of our programs, we expect to have enough work to ensure full employment by 2030, but the shift in automation and artificial intelligence will be important. Career mixes will change and skills and education requirements will change. Redesign work is needed to ensure that humans work most efficiently with the machine.
Workers will need different skills to thrive in the future workplace
Automation will accelerate the shift in the required workforce skills we have witnessed over the past 15 years. The demand for advanced technical skills such as programming will grow rapidly. Social, emotional, and advanced cognitive skills, such as creativity, critical thinking, and complex information processing, will also generate growing demand. The basic demand for digital skills has been increasing and this trend will continue and accelerate. The demand for physical and manual skills will decline, but in many countries, physical machine manual skills will remain the largest workforce skill category in 2030 (see Exhibit 3). This will put additional pressure on existing labor skill challenges and the need for a new qualification system. While some innovative solutions are emerging, there is still a need for solutions that match the scale of the challenge.
Many workers may need to change careers
Our research shows that in the mid-range program, approximately 3% of the global workforce will need to change the occupational category by 2030, although the range of options ranges from 0% to 14%. Some of these changes will occur within companies and departments, but many changes will occur between departments and even regions. A career consisting of a highly structured environment or physical activity of data processing or collection will experience a recession. Growing careers will include activities that are difficult to automate (such as managers) and those who are in an unpredictable physical environment (such as plumbers). Other jobs with increased career needs include teachers, nursing assistants, technology and other professionals
Workplaces and workflows change as more people work with machines
As smart machines and software get more deeply integrated into the workplace, workflows and workspaces will continue to evolve to enable people and machines to work together. For example, as the self-checkout machine is introduced into the store, the cashier can become a checkout assistance assistant who can help answer questions or troubleshoot the machine. More system-level solutions will motivate rethinking the entire workflow and workspace. Warehouse design can change significantly because some parts are designed primarily for robots while others are used to promote safe human-computer interaction.
Automation may put pressure on average wages in advanced economies
Changes in the occupational mix may put pressure on wages. Many of the current middle-wage jobs in advanced economies are primarily highly automated activities, such as manufacturing or accounting, which may decline. High-paying jobs will increase significantly, especially for highly skilled medical and technical or other professionals, but most of the jobs created, including teachers and care assistants, are expected to have a lower salary structure. The risk is that automation can exacerbate wage polarization, income inequality, and insufficient income in developed economies over the past decade, triggering social and political tensions.
Faced with these imminent challenges, the labor challenge already exists
Most countries are already faced with the challenge of fully educating and training their workforce to meet the current demands of employers. OECD spending on worker education and training has been declining over the past two decades. The share of workers’ transfers and misplaced aid as a percentage of GDP has also continued to shrink. One lesson of the past decade is that while globalization may benefit economic growth and the people as consumers, the effects of wages and dislocations on workers have not been adequately addressed. Most analyses, including ourselves, suggest that the scale of these issues may grow in the coming decades. We have also seen in the past that large-scale labor transfers have a lasting impact on wages; during the industrial revolution of the 19th century, despite the increase in productivity, British wages remained stagnant for half a century, this phenomenon is called For "The Stop of Engels" (PDF-690KB).
Ten things to solve
When seeking appropriate measures and policies to address these challenges, we should not seek to roll back or slow down the spread of technology. Companies and governments should use automation and artificial intelligence to benefit from enhanced performance and productivity contributions and social benefits. These technologies will create an economic surplus and help society manage the transformation of the workforce. And the focus should be on ways to ensure that the labor transfer is as smooth as possible. This may require an actionable and scalable solution in several key areas:
Ensure strong economic and productivity growth. Strong growth is not the magical answer to all the challenges of automation, but it is a prerequisite for job growth and growing prosperity. Productivity growth is a key factor in economic growth. Therefore, unlocking investments and requirements, as well as automating the contribution of productivity, is critical.
Develop business vitality. Entrepreneurship and faster new business formation can not only increase productivity, but also drive job creation. A dynamic environment for small businesses and a competitive environment for large companies can spur business and foster employment growth in this way. Accelerating the pace of new business formation and the growth and competitiveness of large and small companies requires simpler and evolving regulations, taxes and other incentives.
An evolving education system and a workplace for learning change. Policy makers working with education providers (traditional and non-traditional) and employers can improve basic STEM skills through school systems and improved on-the-job training. Focus on creativity, critical and systematic thinking, as well as adaptability and lifelong learning. This will require a large-scale solution.
Invest in human capital. Reversing the trend of low tides and, in some countries, reducing public investment in worker training is critical. Through tax incentives and other incentives, policymakers can encourage companies to invest in human capital, including job creation, learning and capacity building, and wage growth, similar to incentives for the private sector to invest in other types of capital, including research and development.
Improve the vitality of the labor market. Information signals that match workers' work and qualifications work better in most economies. Digital platforms can also help people find jobs and restore the vitality of the labor market. When more people change jobs, even within the company, the evidence shows that wages are rising. As more jobs and income-generating opportunities emerge (including performance economies), we need to address issues such as the portability of benefits, worker classification, and wage changes.
Redesign work. Workflow design and workspace design need to adapt to a new era of closer collaboration between people and machines. This is both an opportunity and a challenge in creating a safe and productive environment. Organizations are also changing as jobs become more collaborative and companies seek to become more flexible and non-hierarchical.
Rethinking income. If automation (in whole or in part) does lead to a significant drop in employment and/or huge wage pressures, then some ideas can be considered and tested, such as conditional transfers, liquidity support, universal basic income, and adaptive social safety nets. The key is to find economically viable solutions and combine multiple roles at work, including not only providing income, but also providing meaning, purpose and dignity.
Rethink the transitional support and safety nets of affected workers. As workers work at increasingly higher rates of change between different departments, locations, activities and skill requirements, many workers will need to assist with adjustments. Many best practices for converting safety nets can be used and should be adopted and adjusted, while new methods should be considered and tested.
Invest in the drivers of job demand. Governments need to consider increasing their investment in their own interests and helping to meet job requirements (eg, infrastructure, climate change adaptation). These types of work, from building to rewiring buildings and installing solar panels, tend to be medium-wage jobs, most affected by automation.
Apply artificial intelligence and automation safely. Even if we capture the productivity advantages of these fast-growing technologies, we need to actively guard against risks and mitigate any danger. The use of data must always take into account data security, privacy, malicious use, and potential bias issues, and policy makers, technology, and other companies and individuals need to find effective solutions.
Everyone has a job today and will work for everyone tomorrow, even in the future of automation. However, this work will be different, new skills are needed, and the adaptability of the workforce far exceeds what we see. For the upcoming challenges, training and retraining medium-sized workers and the new generation will be a priority. Governments, private sector leaders and innovators need to work together to better coordinate public and private initiatives, including creating appropriate incentives to invest more in human capital. The future of automation and artificial intelligence will be challenging, but we will benefit a lot if we make the most of technology and mitigate the negative impact.
Transfer from: SMT headline
