1、Generative AI and the future of work in AustraliaFebruary 2024AuthorsChris BradleyJules CarriganGurneet Singh Dandona Seckin UngurCopyright 2024 McKinsey&Company.All rights reserved.This publication is not intended to be used as the basis for trading in the shares of any company or for undertaking a

2、ny other complex or significant financial transaction without consulting appropriate professional advisers.No part of this publication may be copied or redistributed in any form without the prior written consent of McKinsey&Company.Executive summary 2Introduction 121.The accelerated capabilities of

3、machines 132.The new landscape of human work 273.The generative AI opportunity in key economic sectors 404.Considerations for employers,governments,and educators 47Methodology brief 57Acknowledgements 59ContentsIn our 2019 report,Australias automation opportunity:Reigniting productivity and inclusiv

4、e income growth,we examined the possible impact of automation on the future of work.Australias economy was at the tail end of a three-decade boom,and losing momentum fastyet the automation wave was on the horizon,bringing with it the possibility of inclusive economic growth,and the potential to lift

5、 productivity.To realize this promise,Australia would need to embrace rapid automation adoption and,through coordinated action,facilitate social inclusion in the process.Half a decade later,a lot has changednot least because of the profound shifts that the COVID-19 pandemic brought to the Australian

6、 economy.Australian labor markets were resilient during the pandemic amid a series of government subsidies and reforms,but in 2023 generative AI(gen AI)emerged as a significant new force with the potential to reshape the future of work.With its advanced natural language capabilities,gen AI could bec

7、ome ubiquitous,embedded into knowledge workers everyday tools.Workers can use gen AI to write code,design products,create marketing content and strategies,streamline operations,analyze legal documents,provide customer service via chatbots,and even accelerate scientific discovery.But task augmentatio

8、n with a human in the loop rather than outright job automation is more likely at this stage of the technologys development.As gen AI continues to evolve through 2030,it could affect a more comprehensive set of work activities,transforming skills demand in Australia.But gen AI,like all forms of autom

9、ation,does not operate in isolation.Other technologies,including robotics and non-gen AI,may also change the workplace,and in many cases be combined with gen AI.Further,other macroeconomic trends are shaping the demand for certain occupations.For example,shifts accelerated by the pandemic appear to

10、be lasting,including increased remote work and virtual meetings,and the growing popularity of e-commerce.More broadly,Australia is seeing an aging population and increased marketization of unpaid domestic and care workalongside ongoing investments to digitize the economy,uplift the higher education

11、system,and strengthen critical infrastructure.These trends,even without automation,have the potential to change the work landscape and create new occupations.The acceleration of gen AI,combined with overlapping macroeconomic trends,prompted us to reexamine automation and the future of work.McKinseys

12、 latest research,presented in this report,aims to reflect what Australias mix of occupations could look like in 2030,along with potential shifts in skills demand and a view of how workers may need to reskill to stay productively employed as they transition to new roles.Executive summary1 Australias

13、automation opportunity:Reigniting productivity and inclusive income growth,McKinsey,March 3,2019.2 Insights from the first six months of JobKeeper,The Australian Government the Treasury,October 11,2021;The economic potential of generative AI:The next productivity frontier,McKinsey,June 14,2023.3“Wor

14、king from home remains popular but less than in 2021,”Australian Bureau of Statistics,December 13,2023;“Digital activity in the Australian economy,202122,”Australian Bureau of Statistics,October 27,2023.4“Digital economy strategy 2022 update released,”Australian Government Department of the Prime Mi

15、nister and Cabinet,March 30,2022;Jason Clare,“Universities Accord media release,”Australian Government Ministers Media Centre,November 18,2022;Rodney Bogaards,“Infrastructure expenditure over the next decade,”Research Paper Series 202122,Parliament of Australia,April 2022;Older Australians,Australia

16、n Institute of Health and Welfare,June 28,2023;McKinsey Global Institute analysis.2Generative AI and the future of work in AustraliaOur research finds that,in areas such as natural language understanding,and social and emotional reasoning,machines could attain median-and top-quartile human-level per

17、formance much faster than experts previously envisaged.But this need not be a disruption to be feared:if Australia and its workers are well prepared,automation adoption could unlock significant opportunities to improve productivity and economic growth,as well as customer and citizen services,while r

18、eweighting the economy toward higher wageslifting living standards for all.1.The accelerated capabilities of machines Gen AI could accelerate the potential effects of automation,transforming the way work gets done across thousands of activities that make up all the jobs in Australias economy.In 2019

19、,the McKinsey Global Institute estimated that 44 percent of Australians time at work could potentially be automated by adopting technology that existed at that time.This assessment of Australias“technical automation potential”was based on a breakdown of more than 800 occupations into about 2,100 con

20、stituent activities,and an assessment of the capabilities needed to perform each activityranging from natural language understanding to logical reasoning,and problem-solving to fine motor skills.In 2023,we revisited the topic to assess how the rapid emergence of gen AI has accelerated machines capab

21、ilities,finding that 62 percent of existing task hours could be automated using todays technology.This is a 19 percentage point acceleration since the emergence of gen AI.Moreover,this potential could rise further to between 79 and 98 percent by 2030,depending on the scenario.But gen AIs potential d

22、oes not guarantee that businesses and workers will adopt automation solutions at the pace of gen AIs development;even when economic incentives for automation deployment exist,it may take time for adoption to spread across the Australian economy.To understand the pace at which Australias automation p

23、otential might be realizedand the resulting impact on businesses and workerswe modeled a series of adoption scenarios which consider five factors applied to an Australia-specific context:technical feasibility;the time it takes to integrate these capabilities into solutions;the costs associated with

24、developing and deploying solutions;economic feasibility;and the potential for social,organizational,and regulatory to influence adoption.With the advent of gen AI,more than one-quarter of all task hours could be automated by 2030 in the midpoint scenario(Exhibit E1).This is an eight percentage point

25、 acceleration when we compare the same midpoint to the pre-gen AI scenarioa shift that is most marked in knowledge work jobs that previously had relatively low automation potential.For example,a financial auditor might achieve considerable productivity gains by using AI-enabled software to ingest an

26、nual and quarterly reports,check for data anomalies,and detect fraud.5 Australias automation opportunity:Reigniting productivity and inclusive income growth,McKinsey,March 3,2019.6 Based on a historical analysis of various technologies,we modeled a range of adoption timelines from eight to 27 years

27、between the beginning of adoption and its plateau,using sigmoidal curves(S-curves).This range implicitly accounts for the many factors that could affect the pace at which adoption occurs,including regulation,levels of investment,and management decision making within companies.7 The economic potentia

28、l of generative AI:The next productivity frontier,McKinsey,June 14,2023.3Generative AI and the future of work in AustraliaAccordingly,the estimated impact of automation varies widely by occupation.By 2030,one-tenth of Australian workers could see more than 40 percent of their task hours automated by

29、 2030 in a midpoint adoption scenario;two-thirds of workers could see 2040 percent of their task hours automated;while one-quarter of the workforce could experience little or no impact from automation.These midpoint adoption levels could occur only if Australia were to ramp up its R&D investment,tec

30、hnology deployment,and skills development.Our modeling suggests that by 2030,occupations in the highest wage quintile may see their automation adoption increase by 1.8 times due to gen AI,compared to 1.2 times for the lowest wage quintile in the midpoint scenario.The expansion of automation potentia

31、l into knowledge work also means that metropolitan areas with a high concentration of white-collar jobs could face almost as much automation adoption as regional and outer-suburban communities.Exhibit E1Generative AI could unlock new use cases and increase Australias projected automation adoption ra

32、te by eight percentage points by 2030.2022007080902046204420422040203820362034202820262024203220301927+8 p.p.Early scenario including gen AIMidpoint scenario excluding gen AIMidpoint scenario including gen AILate scenario including gen AIAutomation adoption,percentage of time s

33、pent on current work activities1Early scenario:aggressive scenario for all key model parameters(technical automation potential,integration timelines,economic feasibility,and technology difusion rates).2Late scenario:parameters are set for later adoption potential.Source:Australian Bureau of Statisti

34、cs;O*NET;McKinsey Global Institute analysis4Generative AI and the future of work in Australia2.The new landscape of human work The impact of automation on Australian businesses and workers may be substantial.But forecasting the future is no easy task;the acceleration of gen AI is just one of many ov

35、erlapping factors reshaping Australias work landscape.This is why we have broadened our analytical lens to map out what could be a significant transformation of the employment mix by 2030.Our analysis identifies which occupations may experience growing demand,which may face decreased demand,and how

36、much gen AI could change underlying work activities.We also explore how these changes could magnify the importance of skill building to support workers transitioning into new roles.Overall,Australia could experience a rise in healthcare,STEM,and managerial roles,accompanied by a decline in opportuni

37、ties within customer service,office support,and production roles.Amid this shift across occupations,we estimate that 1.3 million workers may need to move into different lines of work by 2030.These transitions could be more pronounced among workers earning within the lowest wage quintile and those wi

38、thout bachelors degrees who are,respectively,5.0 and 1.8 times more likely to experience occupational transitions than their counterparts in the highest wage quintile and workers with higher levels of formal education.At the same time,STEM,healthcare,construction,and professional fields could contin

39、ue to add jobs,even as their work activities change.The pressing question is:will gen AI lead to an overall reduction in employment?Our research does not point to widespread job losses,though there is potential for short-term impacts.Throughout history,technological advancements have often triggered

40、 disruptions,yet they have consistently fueled economic and employment expansion in the long run.When we analyze various factors influencing labor demand,we find that many of the occupational categories most susceptible to gen AI might still experience job growth until 2030,though the adoption of th

41、is technology could temper their growth rates(Exhibit E2).Despite automation,investments and structural factors could continue to sustain employment in these categories.For knowledge workers,however,the likely outcome is that gen AI might significantly alter the nature of their work activities.Amid

42、this shift across occupations,we estimate that 1.3 million workers may need to move into different lines of work by 2030.5Generative AI and the future of work in AustraliaExhibit E2While STEM,healthcare,builders,and professional fields continue to add jobs,generative AI could change work activities

43、for all occupations.4045353025202025234567895Health professionalsHealth aides,technicians,and wellnessAgricultureProduction workCustomer serviceand salesOfce supportMechanical installation and repairTransportation servicesBuildersProperty maintenanceCommunityservicesCreatives a

44、nd arts managementBusiness/legalprofessionalsManagersEducator andworkforce trainingSTEMprofessionalsFood servicesChange in labor demand,%Increase in automation adoption driven by gen AI acceleration,percentage points0Automation adoption by 2030 in the midpoint scenario,%Employment,absolut

45、e0.5m1.0mDecreasing labor demand and modest change of work activitiesIncreasing labor demand and modest change of work activitiesIncreasing labor demand and high change of work activitiesEstimated labor demand change and gen AI automation acceleration by occupational group,AUS,2022301Midpoint automa

46、tion adoption is the average of early and late automation adoption scenarios as referenced in The economic potential of generative AI:The next productivity frontier,McKinsey,June 2023.2We consider multiple drivers afecting demand:rising income,aging populations,technology investment,infrastructure i

47、nvestment,rising education levels,energy transitions,marketization of unpaid work,creation of new occupations,automation(including gen AI),increased remote working and virtual meetings,and e-commerce and other virtual transactions.Source:Australian Bureau of Statistics;O*NET;McKinsey Global Institut

48、e analysis6Generative AI and the future of work in AustraliaFrom our analysis,three distinct occupational groups emerge:Resilient and growing occupations include those in science and technology,healthcare,and professional services.These occupations remained in demand during the pandemic,over the 201

49、922 period;in this group,after automation,there could be net demand for 1.5 million additional jobs in 202230.Up to 210,000 occupational transitions may be required in this category.Stalled but rising occupations,such as building,and mechanical installation and repair,saw downturns from 2019 to 2022

50、 related to the pandemic and global supply shortages.However,these occupations are rebounding as demand for infrastructure grows.By 2030,demand could increase by 290,000 jobs,with 200,000 occupational transitions.Disrupted and declining occupations are those that saw low growth or decline from 2019

51、to 2022 and are estimated to continue to shrink,with as many as 850,000 occupation transitions by 2030one-fifth of the current workforce size.Among these are office support roles,and customer-facing roles affected by the shift to e-commerce and the adoption of gen AI for customer support(Exhibit E3)

52、.Exhibit E3Australian labor markets can be divided into three cohorts.201530 trajectoryExpected occupational transitionsOccupationalgroupResilient and growing 49%of total 2022 employment(6.7m)+1.5m employment,20222030210kBusiness and legal professionalsCreatives and arts managementSTEM professionals

53、Community servicesHealth professionals,aides,technicians,and wellness professionalsManagersTransportation services20030200kStalled but rising20%of 2022 total employment(2.7m)+290k employment,20222030Education and workforce trainingAgricultureBuildersMechanical installation and repairPrope

54、rty maintenance20030850kDisrupted and declining31%of 2022 total employment(4.3m)250k employment,20222030Customer service and salesFood servicesOfce supportProduction work20030Occupations with accelerating automation due to gen AI.We expect continuing future growth but signifcan

55、tly reshaped work activities.Note:Does not include new occupations created by automation.Figures may not sum to 100%,due to rounding.12022 through 2030.Transitions are calculated where there is a decline in net demand for an occupation and employees of that some workforce would have to leave for ano

56、ther occupation.Even in categories that are growing overall,employment may decrease in specifc occupations,requiring some workers to fnd new roles.People joining a new occupation are not counted toward transitions to avoid double counting.Source:Australian Bureau of Statistics;O*NET;McKinsey Global

57、Institute analysisAustralian labor markets could be divided into three cohorts.McKinsey&Company8 Infrastructure beyond COVID-19A national study on the impacts of the pandemic on Australia,Infrastructure Australia,December 16,2020.7Generative AI and the future of work in AustraliaPiecing all this tog

58、ether,workers in lower-wage roles,those with lower levels of formal education,and women are disproportionately represented among those who may need to find new types of work by 2030(Exhibit E4).Workers with these profiles may require increased support for successful transitions.Exhibit E4Workers in

59、lower-wage jobs,those with less education,and women are more likely to be displaced and need to change occupations.Q105101520Q2WageMore likely tochange jobsNationalaverage:9%Less likely tochange jobsEducationGenderQ3Lowest to highest wage quintile5Q47Q53Bachelorsor higher6Less thanbachelors11MenWome

60、n1014148Estimated occupational transitions by 2030,%transition of employment within each category,midpoint automation scenario with gen AI acceleration1Average national transition rate=9.1%,calculated as total estimated transitions by 2030 divided by total employment in 2022.2We split occupations in

61、to quintiles according to their wage distributions.Lowest(Q1)AU$56,200;low(Q2)AU$56,200AU$70,500;middle(Q3)=AU$70,500AU$86,300;high(Q4)=AU$86,300AU$106,600;highest(Q5)AU$106,600.These designations are based on 2022 average annual wages associated with each occupation weighted by 2022 employment.Sour

62、ce:Australian Bureau of Statistics;O*NET;McKinsey Global Institute analysismore likely to be displaced and need to change occupations.McKinsey&Company8Generative AI and the future of work in AustraliaAs the skills landscape shifts,increased demand for activities that require certain skill sets prese

63、nts opportunities for upskilling and transitions to better-paid positionswhich could rebalance the economy toward higher-wage jobs.For instance,in STEM professions,which could see an increase of 250,000 new jobs,91 percent of workers are in higher-wage roles,according to our analysis.On balance,anal

64、ysis suggests an increase in higher-wage jobs,but this does not mean that all workers will benefit.Our modeling also reflects growth for some lower-paid occupations,as they could remain critical to the economy.For example,health aides,technicians,and wellness workers could add almost 300,000 jobs to

65、 the economy by 2030one in five of the net jobs addedyet 90 percent of these workers fall into the lowest wage group.To transition successfully to higher-wage jobs,workers may need access to effective reskilling programs that enable capability building in high-demand skills.Skill building may be a c

66、rucial tool to navigate the future work landscape,and the changing mix of skills needed to complete activities on the job.To keep up with the changes,continuous learning could become an integral component of Australian workers career paths.There could be greater demand for tasks involving social and

67、 emotional skills such as leadership and empathy.And activities which require technological skillsengineering,advanced IT and programming,and digital literacy,for examplecould also be in high demand.On the other hand,basic cognitive skills alone,such as foundational reading,writing,and arithmetic pr

68、oficiency,may not be sufficient to ensure employmentthough they could remain core to many work activities.Increasing demand for jobs that require higher levels of formal educationSTEM jobs,and roles for health professionals,and business and legal professionals,for examplemay increase demand for tert

69、iary degrees.The need for bachelors degrees could grow by 17 percent by 2030,while jobs that require tertiary and higher qualifications could make up about 2 percent more of the labor market.Still,despite this trend toward higher education-level requirements,about 60 percent of the labor market may

70、not require a tertiary degree in 2030.3.The gen AI opportunity in key economic sectorsAs gen AI is expected to transform skill requirements differently across occupations,so too might changes vary across sectors.Consider the manager of a bike rental shop.Every week they spend several hours manually

71、setting up a shift roster for the shops employees,taking into account variables such as leave requests sent via text messages,weather,and the fair allocation of hours.If they used gen AI-enabled scheduling software that draws various data sources to generate schedules,and mediates with employees via

72、 a chatbot,this task could be automated and completed within a few minutes,freeing up time that could be spent productively on other tasks.It is important to note that,similar to many other advanced economies,Australias productivity growth has stagnated.While digital technologies have revolutionized

73、 Australians lives,we have not seen commensurate changes in productivity in the past decade.As with previous waves of technological advancement,flow-on effects to productivity are likely to be slow and unevenexcept in the case of sector-specific innovation.9 Intergenerational report 2023Australias f

74、uture to 2063,Commonwealth of Australia,August 24,2023.9Generative AI and the future of work in AustraliaNonetheless,when we attempt to quantify potential gains from task automation for businesses and sectors in Australias economy,we find that automation and gen AI could improve productivity on a ma

75、cro level.According to our analysis,in theory,Australias productivity could increase by an average of between 0.2 and 4.1 percentage points per yearwith gen AI contributing 0.1 to 1.1 percentage points.In practice,the midpoint adoption scenarioa 4.1 percentage point uplift in productivitymay be quit

76、e ambitious,as it would be four times historical productivity levels.While these numbers are not forecasts,they do give a sense of what is possible as a result of technological advancements.Despite headwinds such as an aging population,automation and gen AI offer opportunities.Indeed,if Australia we

77、re to achieve even half of the potential productivity uplift,it could be on track to rekindle the faster economic growth of the post-1990s heyday.Several enabling factors could help achieve this improvement:leaders who prioritize adoption,redesigned processes,effective change management,and strategi

78、es to ensure value is captured from new efficiencies.For the purposes of this research,we have identified three sectors to bring to life how gen AI could transform the future of work in particular industries:In the retail trade,technology has the capability to introduce greater personalization,redef

79、ining the customer experience.Automation could improve inventory,back-office,and supply chain management,while gen AI could augment key functions such as customer service,and marketing and sales.In financial services and insurance,gen AI adoption could reshape the way employees carry out risk assess

80、ments,fraud detection,software development,and customer service.Nearly one-third of task hours in the sector could be automated by 2030,with much of that opportunity coming from gen AI.In the public sector,gen AI could transform activities such as education delivery,interactions with citizens,financ

81、ial analysis,and R&D.In all these areas,it could provide both productivity gains,and improvements in accuracy and quality of service.In the public sector,gen AI could transform activities such as education delivery,interactions with citizens,financial analysis,and R&D.10 We conservatively use the lo

82、w and midpoint scenarios for these productivity numbers,given uncertainty concerning how productivity benefits will be captured.For midpoint,we created two scenarios:a pessimistic scenario in which labor displaced by automation rejoins the workforce at 2022 productivity levels,and a more optimistic

83、scenario in which it rejoins at 2030 productivity levels,net of automation.In both scenarios,we have incorporated labor displaced rejoining in line with the expected 2030 occupational mix.All other projections,such as jobs lost and jobs gained,are based on the midpoint adoption scenario.10Generative

84、 AI and the future of work in Australia4.Considerations for employers,governments,and educatorsThe accelerated march of automation enabled by gen AI could transform workplaces across Australiabut what is not yet clear is the pace at which this transformation will occur,nor which employers and organi

85、zations will lead the charge.It may not be a straightforward task for Australias stakeholders to realize the full benefits of automation and gen AI,and to ensure that the coming transition in occupations and skills is well planned and fair.Three main groups of stakeholders have an opportunity to tak

86、e meaningful action:employers can harness technology to improve productivity,create value,and help their employees navigate the change;governments can consider implementimg policies to unlock the benefits of automation and establish best practices;and educators can better meet the evolving needs of

87、employers and use gen AI to reimagine how they teach.Employers can consider the following questions as they prepare for the evolution of the workforce:How could gen AI affect our competitive advantage and value proposition?Do we have a strategic workforce plan that matches demand and supply with the

88、 capabilities we need?How can we plan to scale and embed the technologies from day one?Governments can help organizations and citizens unlock the benefits of task augmentation and automation,while providing a transparent regulatory framework and supporting those most vulnerable to role transitions.T

89、o do so,leaders in government can consider the following questions:How can we create a simple,balanced regulatory environment?How can governments enable business adoption of automation?How might we drive automation adoption in the public sector?How can we encourage reskilling and provide a safety ne

90、t for those transitioning to new roles?Education institutions,supported in part by governments,may need to meet the evolving needs of employers,as well as the needs of workers who are transitioning between occupations.Educators can consider the following questions:How can we develop a more responsiv

91、e and agile education system?How can we leverage gen AI to improve student outcomes through personalized learning?Gen AI unlocks a future that differs markedly from the present.Some people may fear the development,thinking it will negatively impact the way that we live and work.Others may embrace it

92、,thinking it will enhance productivity and help meet the needs of the planet and its people.Supporting an optimistic outlook,Australias economy has proved robust through challenging times,and shifts from gen AI could unlock benefits for Australiaincluding higher job demand and productivity.However,t

93、his potential may only be realized if employers,governments,and educators are able to adopt the technology in a bold and thoughtful way.Such strategic action could ensure that Australias future generations benefit from the same prosperity that the country has experienced over the past three decades.

94、11Generative AI and the future of work in AustraliaHow will technology reshape Australians working lives?The potential for automation and artificial intelligence(AI)to disrupt jobs and displace workers looms largethese technological advances could be key to unlocking productivity and value.Based on

95、new research,this report sheds light on these uncertainties,charting the future of work in Australia.In Chapter 1,we consider how new capabilities of machines and automation are transforming thousands of work-related activities in the Australian economy.Anticipating that all sectors stand to be affe

96、cted by automation and generative AI(gen AI),we explore what these changes could mean for Australian businesses and workers.In Chapter 2,we examine the potential impacts of automation,gen AI,and other macro trends on the future of work in Australiawhich include a widespread need for job transitions.

97、We also discuss which individuals may be most vulnerable to these shifts in the jobs landscape,exploring skill building and education as means of transitioning to higher-wage jobs.In Chapter 3,we take a deep dive into the Australian retail trade,financial services,and the public sector to reveal the

98、 opportunities that gen AI presentsincluding the potential to strengthen productivity.We also look at the barriers employers in these sectors may need to overcome to make the most of these opportunities.Finally,in Chapter 4,we consider imperatives for employers,governments,and educators to support t

99、he Australian workforce through the changes ahead and unlock the advantages of automation.We suggest critical questions for each of these stakeholder groups to consider as they prepare for the advance of gen AI and the associated workforce evolution.Introduction12Generative AI and the future of work

100、 in AustraliaThe accelerated capabilities of machines 1.More than 50 robots attended the 2023 AI for Good Global Summit.One of thema humanoid robot named Graceis a nursing assistant who recognizes emotions and shows empathy.Another robot,Sophia,a regular conference attendee and talk show guest,atten

101、ded in her role as the first robot Innovation Ambassador for the United Nations Development Programme.Among their various roles,some robots apply AI to overcome climate-related challenges,while others assist people with disabilities.These examples provide a glimpse of the ways in which AIand gen AI

102、in particularhas accelerated the capabilities of machines to meet and surpass human performance across many capabilities(see sidebar“What is generative AI?”).This chapter assesses how gen AI,along with other technologies,could transform the way work gets done,across all the jobs in Australias econom

103、y:the likely result is a fundamental reshaping of many occupations,particularly those that involve knowledge work.It also considers what gen AI means for Australian businesses and workers,and how it could improve productivity and prosperity for employers and workers alike.11“Meet the robots who are

104、making the world a better place,”UN News,July 6,2023.62 percent of hours spent on existing workforce activities could be potentially automated with existing technology.14Generative AI and the future of work in AustraliaHow machines are transforming work In 2019,the McKinsey Global Institute(MGI)esti

105、mated that 44 percent of Australians time at work could be automated by adopting the technology that existed at that time.This assessment of Australias“technical automation potential”was based on a breakdown of more than 800 occupations into about 2,100 constituent activities,and an assessment of th

106、e capabilities needed to perform each activity.In 2023,the topic was revisited to assess how the rapid emergence of gen AI has accelerated machines capabilities,finding that 62 percent of hours spent on existing workforce activities could be potentially automated with existing technology.This potent

107、ial could rise to between 79 and 98 percent by 2030,depending on the scenario(see“Methodology brief”for details of the analytical approach).What is generative AI?In this report,generative AI(gen AI)encompasses applications typically constructed using foundation models,a class of artificial neural ne

108、tworks that have layered structures comparable to the human brains neuron networks.This type of AI is called generative because of its capacity to create new content based on deep learning,or processing patterns and information from large amounts of data.One example is ChatGPT,which uses its data so

109、urced from the internet to respond to questions and create a variety of written content from poetry and stories to essays and computer code.Gen AI tools can also be used to generate graphics.Leonardo.Ai,developed by a North Sydney-based start-up,generates images and videos online from text prompts.I

110、n an interview for this report,cofounder J.J.Fiasson explains that Leonardo.Ai aims to help people to move quickly from idea to visual creation,an exercise which could otherwise take days or weeks.Users can input basic stick figures which the technology outputs as realistic drawingsfor example,a lin

111、e with some green dashes can be rendered into an oak tree.Despite many users interacting with these platforms for curiositys sake alone,gen AI nevertheless has the potential to perform an array of business-related functionsfrom product design to writing music and from customer service chatbots to as

112、sistance with scientific advancements.Still,gen AIs early stage limitations require consideration to avoid unwanted legal,ethical,or plagiarism-related consequences.Human input and assessment,adequate monitoring,and awareness of possible inaccuracies are critical to ensure that gen AI is used respon

113、sibly.1 For more,see“Generative AI and the future of work in America,”McKinsey Global Institute,July 26,2023 and The economic potential of generative AI:The next productivity frontier,McKinsey,June 14,2023.2 James Purtill,“How Australians are using ChatGPT and other generative AI in their everyday l

114、ives,”ABC Science,April 14,2023.3“Leonardo.Ai accelerates global growth,generates 700 million AI images in less than a year on AWS,”Amazon media alert,November 27,2023.4“Generative AI and the future of work in America,”McKinsey Global Institute,July 26,2023;The economic potential of generative AI:Th

115、e next productivity frontier,McKinsey,June 14,2023.12 Australias automation opportunity:Reigniting productivity and inclusive income growth,McKinsey,March 3,2019.15Generative AI and the future of work in AustraliaGen AI has accelerated machines capabilitiesAI could enable machines to attain top-quar

116、tile human-level performance in critical areas such as natural language understanding,and social and emotional reasoning,two decades faster than envisaged in 2017.Previously forecasted by experts to be achieved by 2050,natural language understanding is now projected to reach top-quartile human-level

117、 performance as early as 2025.A similar level of social and emotional reasoning,originally estimated to be achievable by 2048,may now be accomplished by 2030(Exhibit 1).Exhibit 1Technology could meet and surpass human-level performance in many capabilities much sooner than previously thought.Coordin

118、ation with multiple agentsCreativityLogical reasoning and problem-solvingNatural language generationNatural language understandingOutput articulation and presentationGenerating novel patterns and catergoriesSensory perceptionSocial and emotional outputSocial and emotional reasoningSocial and emotion

119、al sensing20030203520402045205020552060206520702075Line represents range of expert estimatesEstimates before gen AI developments(2017),top quartile Estimates since recent gen AI developments(2023),top quartile Technical capabilities,expected period when level of human performance achievab

120、le by technology1Comparison made based on the business-related tasks required from human workers.Source:McKinsey Global Institute analysis16Generative AI and the future of work in AustraliaMachines are taking on knowledge tasks Gen AIs ability to understand and use natural language for a variety of

121、tasks largely explains why automation potential has risen so steeply.Many activities that involve communication,supervision,documentation,or interacting with people now have the potential to be automated by gen AI.The outlook on the future of work has thus shifted accordingly.Automation technologies

122、 tend to have the greatest impact on tasks that require the lowest skill levels.Gen AI has altered this patternit is rapidly encroaching on higher-skill tasks too.Australians spend over 40 percent of their working hours on“knowledge work”that involves decision making and collaboration:applying exper

123、tise,managing people,and interacting with stakeholders(Exhibit 2,part 1).Gen AI has doubled or tripled the automation potential for such activities.For example,55 percent of the time spent on activities that require workers to apply expertise has the potential to be automated.Without gen AI,the shar

124、e drops to 20 percent.As a result,knowledge work occupations such legal professions,computer engineering,and teaching could be most affected by gen AI,with specific activities automation potential varying significantly across occupations(Exhibit 2,part 2).In comparison,predictable physical work,such

125、 as delivering packages or maintaining laboratory equipment,could remain relatively untouched by gen AI.At 67 percent,the overall automation potential for these activities is still higher than that of decision making and collaboration activities.However,excluding gen AI would only reduce the automat

126、ion potential for predictableand unpredictablephysical work by one percentage point(Exhibit 2,part 1).13 Australias automation opportunity:Reigniting productivity and inclusive income growth,McKinsey,March 3,2019.Many activities that involve communication,supervision,documentation,or interacting wit

127、h people now have the potential to be automated by gen AI.17Generative AI and the future of work in AustraliaExhibit 2,part 1Knowledge work,including decision making and collaboration,could be most impacted by generative AI.ActivitygroupsAutomation potential,Comparison in midpoint scenarios,2023,%Sh

128、are oftime spent,%Example activitiesPlan facility layouts or designsConduct fnancial or regulatory auditsDirect employee training programsSupervise employeesPromote products or programsPresent information to publicAnalyze market research dataModel operational processesMaintain compliance documentati

129、onMonitor organizational processesMaintain test equipmentDeliver itemsFabricate devices or componentsCut glassDecision making and collaborationApplying expertiseManagingInterfacing withstakeholdersProcessing dataCollecting dataPerforming predictablephysical workPerforming unpredictablephysical workD

130、ata managementPhysical20+35 p.p.5514+29 p.p.424923+25 p.p.74+17 p.p.9159+17 p.p.7666+1 p.p.673534+1 p.p.10With gen AIWithout gen AIAutomation potential by type of work activity,2023,%Note:Figures may not sum,because of rounding.1Applying expertise to decision making,planning,and creative

131、tasks.2Managing and developing people.Source:Australian Bureau of Statistics;O*NET;McKinsey Global Institute analysis18Generative AI and the future of work in AustraliaExhibit 2,part 2The estimated impact of automation varies widely by occupation.ActivitygroupsDecision making and collaboration:Apply

132、ing expertiseManagingInterfacing withstakeholdersProcessing dataCollecting dataPerforming physical activities and operating machinery in predictable environmentsPerforming physical activities and operating machinery in unpredictable environmentsData management:Physical:Architects,except landscape an

133、d navalCashiersGraphic designersHairdressers,hairstylists,and cosmetologistsJudicial law clerksParalegals and legal assistantsLow(66%)Examples of automation potential by type of work activity,2023,%The estimated impact of automation varies widely by occupation.McKinsey&Company1A blank space represen

134、ts instances where the occupation does not typically perform the corresponding activity.2Applying expertise to decision making,planning,and creative tasks.3Managing and developing people.Source:Australian Bureau of Statistics;O*NET;McKinsey Global Institute analysis19Generative AI and the future of

135、work in AustraliaFrom potential to adoption:What gen AI means for Australian businesses and workersAlthough 62 percent of the Australian workforces activities are potentially automatable,the actual pace of adoption usually lags behind technical potential.Technological capabilities may be available b

136、ut not yet integrated into a solution that automates a specific work activitydeveloping such solutions takes time.Even when a solution is developed,it might not be economically feasible to use if its costs exceed those of human labor.Additionally,even when economic incentives exist,it may take time

137、for adoption to spread across the Australian economy.For instance,the early technology required for self-driving cars exists today,but(in contrast to some overseas jurisdictions)few jurisdictions in Australia have made them legal to use.We modeled a series of adoption scenarios that consider five fa

138、ctors that affect the pace and extent of automation:1.Technical feasibilityif capabilities can automate a particular activity.2.The time it takes to integrate these capabilities into solutions.3.Costs associated with developing and deploying solutions.4.Economic feasibility,contingent on labor marke

139、t dynamics(wages).5.Other factors that have an impact on adoption,both regulatory and social,including pace of organizational change,regulatory choices,and stakeholder acceptance.Accounting for all of these sources of friction,there could be a substantial time lag between technical potential and rea

140、lized change.Therefore,while the early scenario suggests just above 50 percent of activities could be automated by 2030,the late scenario could see just 2 percent in the same year.The midpoint of these scenarios would imply that around one-quarter of work hours could be automated by 2030.This is an

141、eight percentage point acceleration with the inclusion of gen AI(Exhibit 3).14 The regulatory framework for automated vehicles in Australia:Policy paper,National Transport Commision,Melbourne,February 2022.15 Based on a historical analysis of various technologies,we modeled a range of adoption timel

142、ines from eight to 27 years between the beginning of adoption and its plateau,using sigmoidal curves(S-curves).This range implicitly accounts for the many factors that could affect the pace at which adoption occurs,including regulation,levels of investment,and management decision making within compa

143、nies.The midpoint of these scenarios would imply that around one-quarter of work hours could be automated by 2030.This is an eight percentage point acceleration with the inclusion of gen AI.20Generative AI and the future of work in AustraliaEven for midpoint adoption to occur,Australia would need to

144、 ramp up its R&D investment,technical development potential,and technology deployment.Regulatory,organizational,and social factors would also require consideration.If Australia and its businesses do not overcome the significant hurdles,the country could find itself in the late scenario,with very lim

145、ited technology adoption and low international competitiveness.All sectors may be affected by automation and gen AIIn the midpoint adoption scenario,every sector and occupation in Australia sees increases in automation,with associated potential for productivity gains.Education,professional,scientifi

146、c,and technical services,and finance and insurance are anticipated to experience the largest increases in opportunity for adoption by 2030(Exhibit 4).Exhibit 3Generative AI could unlock new use cases and increase Australias projected automation adoption rate by eight percentage points by 2030.202201

147、020304050607080902046204420422040203820362034202820262024203220301927+8 p.p.Early scenario including gen AIMidpoint scenario excluding gen AIMidpoint scenario including gen AILate scenario including gen AIAutomation adoption,percentage of time spent on current work activities1Early scenar

148、io:aggressive scenario for all key model parameters(technical automation potential,integration timelines,economic feasibility,and technology difusion rates).2Late scenario:parameters are set for later adoption potential.Source:Australian Bureau of Statistics;O*NET;McKinsey Global Institute analysis2

149、1Generative AI and the future of work in AustraliaWhite-collar workers could be disproportionately affected by gen AI Gen AI is likely to have the largest incremental impact on white-collar jobs.Teachers,in particular,could see their use of time transformed by gen AI technologies.For example,gen AI

150、could augment the assignment-marking process,making more room for student-facing teaching.For workers in the engineering sector,gen AI could accelerate the design process and reduce errors and reworking,simplifying complex mechanical,electrical,and plumbing systemswhile taking local council regulati

151、ons into account.Gen AI could also empower engineers to consider energy efficiency,carbon footprint,and material use.This means that at a time of workforce shortages and supply chain issues in the construction sector,Australia could deliver affordable housing and major infrastructure projects more e

152、ffectively.Exhibit 4Education,professional,and financial services could see the largest increases in automation opportunity from generative AI by 2030.Sector%of time spentEducational servicesProfessional,scientifc,and technical servicesFinance and insuranceInformationAdministrative and support and g

153、overnmentReal estate and rental and leasingUtilitiesArts,entertainment,and recreationMiningWholesale tradeConstructionTransportation and warehousingManufacturingOther servicesAccommodation and food servicesHealthcare and social assistanceAgriculture,forestry,fshing,and huntingRetail trade

154、886583327282673322728256322527Automation adoption by 2030,excluding gen AIMost impacted sectors from gen AIAutomation adoption by 2030,incremental from gen AIProjected automation adoption by sector,midpoint automation scenario,2030,%Note:Does no

155、t include new occupations created by automation.Figures may not sum,because of rounding.1Midpoint of earliest and latest automation adoption with gen AI.Source:Australian Bureau of Statistics;O*NET;McKinsey Global Institute analysis22Generative AI and the future of work in AustraliaGen AIs reconfigu

156、ration of white-collar jobs may also be felt in higher-wage roles.Australians earning within the highest wage quintile are expected to experience the largest increase in adoption of automationan increase of 1.8 times compared to 1.2 times for the lowest wage quintile.This marks a departure from auto

157、mations disproportionate association with lower-wage professions.Projections indicate a 26 percent automation rate for higher-wage jobs by 2030,closely trailing the 28 percent estimate for lower-wage jobs.Gen AI extends automations reach into city centersAll regions of Australia could feel the effec

158、ts of automation and gen AI.At state and territory levels,gen AI could add about eight percentage points to each states levels of automation adoption,narrowing the gap between the highest and lowest state levels of automation.The Australian Capital Territory,with a high concentration of white-collar

159、 jobs,moves from least automatable(17.7 percent)to most(28.0 percent).Western Australia,with a large mining industry,has the highest level of non-gen AI automation(19.7 percent)and sees an 8.2 percentage point increase in adoption of automation just from gen AI.Our 2019 research showed the highest l

160、evels of automation in outer suburban and regional communities,with sectors particularly vulnerable to automation but without the skills and resilience base to adapt as quickly.Conversely,regions with high concentrations of white-collar jobs,such as professional services,would have been the least im

161、pacted by automation.This story has now become more nuanced as gen AI balances out some of these differences.When examining Statistical Area Level 4(SA4)regions,metropolitan areas,with a high concentration of white-collar jobs,are now facing almost as much automation adoption as regional and outer s

162、uburban communities(Exhibit 5).All regions of Australia could feel the effects of automation and gen AI.At state and territory levels,gen AI could add about eight percentage points to each states levels of automation adoption,narrowing the gap between the highest and lowest state levels of automatio

163、n.16 Australias automation opportunity:Reigniting productivity and inclusive income growth,McKinsey,March 3,2019.17 Ibid.18 Ibid.23Generative AI and the future of work in AustraliaExhibit 5Generative AIs impact is ubiquitous and creates a more balanced automation distribution.6.3Automation adoption

164、by 2030,incremental from gen AI,percentage points11.2128910MackayIsaacWhitsunday22%automation rate without gen AI(10th decile)30%automation rate with gen AI(10th decile)Top 3 employment sectors(%of total jobs):mining(16%),healthcare and social assistance(11%),and retail trade(9%)2Western AustraliaOu

165、tback(North)22%automation rate without gen AI(10th decile)31%automation rate with gen AI(10th decile)Top 3 employment sectors(%of total jobs):mining(26%),healthcare and social assistance(11%),and administrative and support,and government(10%)1SydneyCity and Inner South17%automation rate without gen

166、AI(1st decile)28%automation rate with gen AI(9th decile)Top 3 employment sectors(%of total jobs):professional,scientifc,and technical services(19%),healthcare and social assistance(10%),and fnance and insurance(10%)4SydneyParramatta19%automation rate without gen AI(6th decile)28%automation rate with

167、 gen AI(10th decile)Top 3 employment sectors (%of total jobs):healthcare and social assistance(14%),professional,scientifc,and technical services(12%),and retail trade(10%)36 MelbourneWest20%automation rate without gen AI(7th decile)28%automation rate with gen AI(8th decile)Top 3 employment sectors(

168、%of total jobs):healthcare and social assistance(12%),transportation and warehousing(10%),and retail trade(10%)MelbourneInner17%automation rate without gen AI(1st decile)27%automation rate with gen AI(7th decile)Top 3 employment sectors(%of total jobs):professional,scientifc,and technical services(1

169、9%),healthcare and social assistance(13%),and administrative and support,and government(9%)5BarossaYorke(Mid North)20%automation rate without gen AI(8th decile)27%automation rate with gen AI(4th decile)Top 3 employment sectors(%of total jobs):healthcare and social assistance(14%),agriculture,forestr

170、y,fshing and hunting(12%),and manufacturing(12%)8Geelong19%automation rate without gen AI(3rd decile)26%automation rate with gen AI(2nd decile)Top 3 employment sectors(%of total jobs):healthcare and social assistance(17%),construction(12%),and retail trade(11%)7PerthInner17%automation rate without g

171、en AI(1st decile)27%automation rate with gen AI(7th decile)Top 3 employment sectors(%of total jobs):professional,scientifc,and technical services(17%),healthcare and social assistance(17%),mining(9%)10Bunbury21%automation rate without gen AI(10th decile)28%automation rate with gen AI(9th decile)Top

172、3 employment sectors(%of total jobs):healthcare and social assistance(13%),construction(11%),and retail trade(10%)9Incremental growth in automation from gen AI by SA4 region,percentage points Note:The boundaries and names shown on this map do not imply ofcial endorsement or acceptance by

173、McKinsey&Company.Source:Australian Bureau of Statistics;O*NET;McKinsey Global Institute analysis24Generative AI and the future of work in AustraliaThe narrowing gap between different regions because of gen AIs impact on white-collar jobs is exemplified by the comparison between Inner Sydney and subu

174、rbs around Parramatta.Parramatta has a higher proportion of jobs that comprise tasks automatable without gen AI,and Inner Sydney has high concentrations of occupations where automatable tasks could increase significantly with the adoption of gen AI.Gen AI closes the automation gap between Parramatta

175、 and Inner Sydney from 2.4 percentage points without gen AI to only 0.2 percentage points with gen AI(Exhibit 6).Exhibit 6Major metro areas are most impacted by generative AI due to a high concentration of white-collar jobs.SydneyNorth Sydney and HornsbyTotal jobs:254,00016%automation rate without g

176、en AI(1st decile)27%automation rate with gen AI(7th decile)Top 3 employment sectors(%of total jobs):professional,scientifc,and technical services(20%),healthcare and social assistance(14%),and fnance and insurance(13%)SydneyCity and Inner SouthTotal jobs:224,00017%automation rate without gen AI(1st

177、decile)28%automation rate with gen AI(9th decile)Top 3 employment sectors(%of total jobs):professional,scientifc,and technical services(19%),healthcare and social assistance(10%),and fnance and insurance(10%)SydneyParramattaTotal jobs:230,00019%automation rate without gen AI(6th decile)28%automation

178、 rate with gen AI(10th decile)Top 3 employment sectors(%of total jobs):healthcare and social assistance(14%),professional,scientifc,and technical services(12%),and retail trade(10%)6.3%Automation adoption by 2030,incremental from gen AI,percentage points11.2%Incremental growth in automation from gen

179、 AI by SA4 region,percentage points Note:The boundaries and names shown on this map do not imply ofcial endorsement or acceptance by McKinsey&Company.Source:Australian Bureau of Statistics;O*NET;McKinsey Global Institute analysis19 Inner Sydney corresponds to SA4 region SydneyCity and Inner South;Pa

180、rramatta corresponds to SA4 region SydneyParramatta.25Generative AI and the future of work in AustraliaThe same trend can be observed when comparing the inner city and regional areas of Melbourne and Perth,respectively.Despite this rebalancing of automations impact driven by gen AI,automatability st

181、ill increases in every region.As we discuss in Chapter 2,in the midpoint adoption scenario,new skills may be required across all occupations.This can be more difficult in outer-suburban and regional areas,as formal and informal education may be more limited.But with edtech,this gap can be bridged if

182、 internet penetration and digitization increase,and knowledge can be transferred.Regions with high concentrations of white-collar jobs,such as professional services,would have been the least impacted by automation.This story has now become more nuanced as gen AI balances out some of these difference

183、s.26Generative AI and the future of work in AustraliaThe new landscape of human work2.The impact of automation on Australian businesses and workers may be substantial.But the acceleration of gen AI is only one of many overlapping trends reshaping Australias work landscape.Therefore,this report uses

184、a broader lens,building on previous MGI research on the future of work.It considers,for example,that the pandemic prompted a steep rise in the number of people working remotely and shopping online.It also takes into account other technologies including robotics and non-gen AI,Australias aging popula

185、tion and increased marketization of unpaid domestic and care work,and Australias ongoing investments to digitize the economy,uplift the higher education system,and strengthen critical infrastructuretrends that,even without automation,could result in the creation of new occupations.This chapter first

186、 explores the combined impact of automation,gen AI,and broader macro trends on the future demand for occupations and the resulting widespread need for job transitions.It then determines which individuals could be disproportionately affected by the shifting jobs landscape,requiring reskilling support

187、 to weather the transition.Forecasting the future is no easy task.However,we can anticipate how various trends may impact the mix of jobs in Australias economy by the end of the decade.Our analysis distinguishes between occupations which may experience growing demand and those which may face job los

188、ses,and considers how much gen AI can change underlying work activities.We find that job growth may still persist in occupational categories most vulnerable to gen AI through 2030,even though the adoption of this technology might moderate their growth rates.Knowledge workers could expect a definite

189、transformation in the nature of their work activities,as gen AI takes a substantial role in shaping their professional landscape.For some categories,such as office support,customer service,and production work,automation could contribute to workforce reductions(Exhibit 7).Overall,an estimated 1.3 mil

190、lion occupational transitions could occur in Australia through 2030.Understanding the nuances of these changes,and their potential impact on individuals and businesses,is crucial for a smoother transition.For instance,roles that are categorized within the lowest wage quintile and those without bache

191、lors degree requirements are,respectively,5.0 and 1.8 times more likely to experience occupational transitions than roles within the highest wage quintile and with higher education requirements.20 See the following McKinsey and MGI reports:The economic potential of generative AI:The next productivit

192、y frontier,June 14,2023;Jobs lost,jobs gained:Workforce transitions in a time of automation,December 2017;The future of work in America:People and places,today and tomorrow,July 11,2019;and The future of work after COVID-19,February 18,2021;Generative AI and the future of work in America,July 26,202

193、3;Australias automation opportunity:Reigniting productivity and inclusive income growth,McKinsey,March 3,2019.21“Working from home remains popular but less than in 2021,”Australian Bureau of Statistics,December 13,2023;“Digital activity in the Australian economy,202122,”Australian Bureau of Statisti

194、cs,October 27,2023.22“Digital economy strategy 2022 update released,”Australian Government Department of the Prime Minister and Cabinet,March 30,2022;Jason Clare,“Universities Accord media release,”Australian Government Ministers Media Centre,November 18,2022;Rodney Bogaards,“Infrastructure expendit

195、ure over the next decade,”Research Paper Series 202122,Parliament of Australia,April 2022;Older Australians,Australian Institute of Health and Welfare,June 28,2023;McKinsey Global Institute analysis.28Generative AI and the future of work in AustraliaExhibit 7While STEM,healthcare,builders,and profes

196、sional fields continue to add jobs,generative AI could change work activities for all occupations.4045353025202025234567895Health professionalsHealth aides,technicians,and wellnessAgricultureProduction workCustomer serviceand salesOfce supportMechanical installation and repairT

197、ransportation servicesBuildersProperty maintenanceCommunityservicesCreatives and arts managementBusiness/legalprofessionalsManagersEducator andworkforce trainingSTEMprofessionalsFood servicesChange in labor demand,%Increase in automation adoption driven by gen AI acceleration,percentage points152525

198、353540Automation adoption by 2030 in the midpoint scenario,%Employment,absolute0.5m1.0mDecreasing labor demand and modest change of work activitiesIncreasing labor demand and modest change of work activitiesIncreasing labor demand and high change of work activitiesEstimated labor demand change and g

199、en AI automation acceleration by occupational group,AUS,2022301Midpoint automation adoption is the average of early and late automation adoption scenarios as referenced in The economic potential of generative AI:The next productivity frontier,McKinsey,June 2023.2We consider multiple drivers afecting

200、 demand:rising income,aging populations,technology investment,infrastructure investment,rising education levels,energy transitions,marketization of unpaid work,creation of new occupations,automation(including gen AI),increased remote working and virtual meetings,and e-commerce and other virtual tran

201、sactions.Source:Australian Bureau of Statistics;O*NET;McKinsey Global Institute analysis29Generative AI and the future of work in AustraliaExhibit 8Healthcare,STEM,business and legal professionals,and managers are likely to comprise about 85 percent of net job growth by 2030.Net change 202230 FTE,th

202、ousands7375358648302993671,600Occupational groupTotalManagersHealth aides,technicians,and wellnessSTEM professionalsHealth professionalsBusiness/legal professionalsBuildersTransportation servicesEducator and workforce trainingCommunity servicesCreatives and arts managementPrope

203、rty maintenanceMechanical installation and repairAgricultureFood servicesProduction workCustomer service and salesOfce supportNet change in FTE 202230,CAGR,%1.40.60.30.10.41.11.21.81.40.91.92.11.73.73.23.82.61.485%Projected net change in jobs by occupation type,midpoint automation scenario,with gen

204、AI acceleration,202230,thousands Note:Does not include new occupations created by automation.Figures may not sum,because of rounding.Source:Australian Bureau of Statistics;O*NET;McKinsey Global Institute analysisThe changing demand for workers could reshape the Australian workforce by 2030Based on t

205、he modeling,we could expect Australias employment mix to change significantly through 2030.Certain occupational categoriesparticularly in healthcare,science and technology,and other high-skill professional servicesare likely to experience growth.At the same time,demand could decrease for occupations

206、 such as customer and office support:around 190,000 fewer office support roles may be needed come 2030,driven by automation and evolving worker preferences,including increased remote work(Exhibit 8).30Generative AI and the future of work in AustraliaWhen looking at future demand for jobs,three disti

207、nct occupational groups emerge:resilient and growing,stalled but rising,and disrupted and declining.Resilient and growing occupations include those in STEM,healthcare,and professional services.There could be a demand for 1.5 million new jobs in this category,including 245,000 health professionals an

208、d 300,000 health aides,health technicians,and wellness workersroles associated with the specialized healthcare requirements of Australias aging population.As the service-based nature of the Australian economy continues to strengthen,demand for 365,000 new managers and over 180,000 new businesses and

209、 legal professionals can also be expected.Alongside trends such as digitization and e-commerce,STEM jobs are estimated to increase by 29 percent,equivalent to an average growth rate of 3.2 percent per year through 2030adding a quarter of a million jobs.The rise of e-commerce may elevate demand for t

210、ransportation services too,adding over 80,000 more jobs as consumers increasingly shop online and rely on delivery services to obtain their goods.Stalled but rising occupations,such as building and mechanical installation and repair,saw downturns from 2019 to 2022 related to the pandemic and global

211、supply shortages.However,these occupations,comprising over 60 percent of increases in this category,are rebounding as demand for infrastructure grows.By 2030,demand is estimated to increase by 290,000 jobs,while education and workforce training could add 64,000 more jobs.Disrupted and declining occu

212、pations are those that grew slowly or declined from 2019 to 2022;they are likely to continue to shrink.Among these are customer-facing roles,affected by the shift to e-commerce,and office support roles,likely affected by automation or decreased on-site office work.Declines in office support,producti

213、on work,and customer service and sales could account for 260,000 job losses by 2030.These jobs involve a high share of repetitive tasks,data collection,and basic data processing,all activities that automated systems can undertake efficiently.Demand shifts could result in 1.3 million job transitionsT

214、echnology-driven shifts intersect with other macro factors such as an aging population,the net-zero transition,and increased infrastructure spending.When these factors are combined,up to 1.3 million workers9 percent of Australias total workforcemay need to transition out of their current roles into

215、new occupations by 2030(Exhibit 9).The need to transition arises when an individual is involuntarily displaced from a job and unable to find a role in the same occupation elsewhere due to trends such as automation,a shift to e-commerce,or other changing business models.As a result,these individuals

216、find themselves unable to secure a new job in the same occupation due to declining demand.23 Our STEM skilled futureAn education roadmap for an innovative workforce,Australian Academy of Technological Sciences and Engineering,October 2022.24 Infrastructure beyond COVID-19A national study on the impa

217、cts of the pandemic on Australia,Infrastructure Australia,December 16,2020.25 Here,1.3 million transitions are distinct from regular employment churn within the economy.In this case,an individual would move into a different occupation from their current one,as opposed to current measures of churn wh

218、ere an individual may leave one business and go to another business to perform the same occupation.31Generative AI and the future of work in AustraliaExhibit 9Australian labor markets could be divided into three cohorts.201530 trajectoryExpected occupational transitionsOccupationalgroupResilient and

219、 growing 49%of total 2022 employment(6.7m)+1.5m employment,20222030210kBusiness and legal professionalsCreatives and arts managementSTEM professionalsCommunity servicesHealth professionals,aides,technicians,and wellness professionalsManagersTransportation services20030200kStalled but risi

220、ng20%of 2022 total employment(2.7m)+290k employment,20222030Education and workforce trainingAgricultureBuildersMechanical installation and repairProperty maintenance20030850kDisrupted and declining31%of 2022 total employment(4.3m)250k employment,20222030Customer service and salesFood serv

221、icesOfce supportProduction work20030Occupations with accelerating automation due to gen AI.We expect continuing future growth but signifcantly reshaped work activities.Note:Does not include new occupations created by automation.Figures may not sum to 100%,due to rounding.12022 through 203

222、0.Transitions are calculated where there is a decline in net demand for an occupation and employees of that some workforce would have to leave for another occupation.Even in categories that are growing overall,employment may decrease in specifc occupations,requiring some workers to fnd new roles.Peo

223、ple joining a new occupation are not counted toward transitions to avoid double counting.Source:Australian Bureau of Statistics;O*NET;McKinsey Global Institute analysisAustralian labor markets could be divided into three cohorts.McKinsey&CompanySome displaced workers may redirect their skills and ex

224、pertise toward alternative jobs within the same occupational category.For instance,an accounting clerk could retrain as a database administrator to stay employed within the office support category.Other workers may need to change roles outright.A call center agent displaced by the decline in the cus

225、tomer service category,for example,may retrain as a medical assistant,where there is growing demand.The need for job transitions can be categorized into the same three occupational groups used to categorize changes in job demand:resilient and growing,stalled but rising,and disrupted and declining(Ex

226、hibit 9).32Generative AI and the future of work in AustraliaExhibit 101.3 million occupational transitions could occur by 2030.Occupational groupOccupational transitions churn rate,%of 2022 employment1.3 million(9%of all jobs)13.8 millionOccupation transitions,202230,thousandsEmployment,2022,thousan

227、dsShare of workers with bachelors degree or above,2022,%Health professionals0373590TotalHealth aides,technicians,and wellness1985328Creatives and arts management41235350STEM professionals21886774Property maintenance62336414Transportation services52551615Managers2271,63245Community services62950842Ag

228、riculture103232115Mechanical installation and repair11454107Educator and workforce training54683677Builders6507975Business/legal professionals7891,27466Food services1311485414Production work1913874310Customer service and sales1817094718Ofce support24 429 1,79525Resilient and continuing to growStalle

229、d but starting to riseDisrupted and continuing to declineEstimated number of occupational transitions by occupational group,202230,midpoint automation scenario with gen AI acceleration1Churn rate is calculated as occupation transitions divided by 2022 employment.Source:Australian Bureau of Statistic

230、s;O*NET;McKinsey Global Institute analysisThe resilient and growing and stalled but rising categories could see up to 210,000 and 200,000 occupation transitions by 2030,respectively.Despite high growth overall,these two groups account for just over 30 percent of all transitions.The disrupted and dec

231、lining category,which could experience a net decline in jobs,might account for more than 65 percent of the total 1.3 million estimated occupational transitions by 2030.Declining demand for jobs in office support,production work,food services,and customer service and sales could see almost 850,000 wo

232、rkers leaving their current occupations and finding jobs in different occupations(Exhibit 10).33Generative AI and the future of work in AustraliaExhibit 11Workers in lower-wage jobs,those with less education,and women are more likely to be displaced and need to change occupations.Q105101520Q2WageMor

233、e likely tochange jobsNationalaverage:9%Less likely tochange jobsEducationGenderQ3Lowest to highest wage quintile5Q47Q53Bachelorsor higher6Less thanbachelors11MenWomen1014148Estimated occupational transitions by 2030,%transition of employment within each category,midpoint automation scenario with ge

234、n AI acceleration1Average national transition rate=9.1%,calculated as total estimated transitions by 2030 divided by total employment in 2022.2We split occupations into quintiles according to their wage distributions.Lowest(Q1)AU$91kMiddle-wageAU$66kAU$91kLow-wageAU$66k 2.3 0.52.8 9252220

235、1815116418 65899626858434699086100 STEM professionalsManagersBusiness/legal professionalsCommunity servicesEducation and workforce trainingMechanical installation and repairBuildersProduction workHealth professionalsCreatives and arts managementAgricultureCustomer se

236、rvice and salesTransportation servicesOfce supportHealth aides,technicians,and wellnessProperty maintenanceFood services High wage Middle wage Low wage Note:Figures may not sum to 100%,because of rounding.1Based on 2022 wages.2Lower-wage jobs are those below the 30th percentile;mid-wage jobs are tho

237、se between the 30th and 70th percentiles;and high-wage jobs are those above the 70th percentile of wage distribution.Growth from 2022 to 2030 holds 2022 wage categorization constant.Source:Australian Bureau of Statistics;O*NET;McKinsey Global Institute analysis38Generative AI and the future of work

238、in AustraliaExhibit 14To move to higher-wage jobs of the future,workers are likely to need more social and emotional and technological skills.76224024271711Physical and manualBasic cognitiveHigher cognitiveSocial and emotionalTechnologicalQ1Q2Q3Q4Q5Highest wage quint

239、ileLowest wage quintileTime spent using various types of skills by wage quintile,1 2030,%Note:Figures may not sum to 100%,because of rounding.1Using O*NET data,we classifed 2,100 work activities associated with 850 occupations according to the primary type of skill used.Source:Australian Bureau of S

240、tatistics;O*NET;McKinsey Global Institute analysis39Generative AI and the future of work in AustraliaThe generative AI opportunity in key economic sectors3.The manager at a bicycle rental store spends several hours manually compiling the weekly shift scheduleconsulting weather forecasts and past sal

241、es reports to predict busy days,juggling time-off requests that employees send via text,and making sure that the hour allocation is fair.With gen AI-powered software,this weekly requirement could be reduced to a ten-minute task,freeing up time to mentor staff or plan new offerings for the store.Empl

242、oyees would have access to a chatbot that logs their requests and sends reminders or updates,and the program would analyze historical sales data,store policies,weather patterns,holidays,and local labor regulations to generate dynamic shift schedules that adjust in real time as variables fluctuate.As

243、 gen AI is expected to transform skill requirements differently across occupations,so too might changes vary across sectors.In this chapter,we consider the potential for gen AI to strengthen productivity across the Australian economyand the barriers that would need to be overcome to realize that pot

244、ential.We then take a closer look at the specific opportunities that gen AI presents to employers in three key economic sectors:the retail trade,financial services,and the public sector.Successful integration of automation and gen AI could boost the Australian economy and benefit businesses After pe

245、aking at a 2 percent annualized growth rate during the 19922002 decade,productivity growth in Australia halved in the two decades that followed,touching 1 percent(Exhibit 15).The slowdown in productivity and stagnating employment growth have led to a downturn in the countrys overall GDP growth.With

246、aging contributing to relatively lower growth in the labor force in the future,productivity growth acceleration is imperative to achieve the higher GDP growth observed in previous decades.Gen AI can boost productivity for individual tasks,but the pace of its overall effect on the economy is uncertai

247、n.To have an impact on productivity,technologies would need to be implemented in the“real world”to automate and augment a wide range of tasks,and the time saved would need to be redeployed back into productive work.In the bicycle shop example,the AI technology could reduce the time spent on manual r

248、ostering,and the manager would need to spend that time on other value-additive activities.If task automation gains flow through to many businesses across sectors of Australias economy,then automation and gen AI could improve productivity on a macro level.But this may require the right enabling lever

249、s to be in placesuch as leaders being willing to prioritize adoption,redesigning processes to use the technology effectively,change management to get employees comfortable with the technology,and strategies to ensure value is captured from the efficiencies gained.Assuming all factors are in place,ou

250、r modeling suggests that gen AI has the potential to increase Australian labor productivity by 0.1 to 1.1 percentage points a year through 2030.The range reflects a late and average speed of gen AI adoption,along with the time of full-time employees released from deploying these technologies being r

251、edeployed back into the economy.Both scenarios account 30 Intergenerational report 2023:Australias future to 2063,Australian Government,2023.31 We conservatively use the low and midpoint scenarios for these productivity numbers,given uncertainty concerning how productivity benefits will be captured.

252、For both low and midpoint,we created two scenarios:a pessimistic scenario in which labor displaced by automation rejoins the workforce at 2022 productivity levels,and a more optimistic scenario in which it rejoins at 2030 productivity levels,net of automation.In both scenarios,we have incorporated l

253、abor displaced rejoining in line with the expected 2030 occupational mix.All other projections(such as,for example,jobs lost and jobs gained),are based on the midpoint adoption scenario.41Generative AI and the future of work in Australiafor the occupational mix expected in 2030.When we combine gen A

254、I with all other automation technologies,the productivity growth could range from 0.2 to 4.1 percent a year in the late and midpoint adoption scenarios,respectively.But as a year-on-year increase of 4.1 percentage points would be up to four times greater than recent historical productivity growth le

255、vels,this potential is unlikely to be fully realized,especially as there could be significant transition costs and second-order effects(Exhibit 15).Exhibit 15Automation adoption and redeploying full-time equivalent hours back into the workforce could boost Australias stalled productivity growth.Note

256、:Figures may not sum,because of rounding.Source:The Conference Board Total Economy database;McKinsey Global Institute analysis McKinsey&CompanyAustralia productivity growth,compound annual growth rate,%Automation adoption without genAIAdditional with gen AIAssuming displaced FTEs rejoin at 2022 prod

257、uctivity ratesAssuming displaced FTEs rejoin at 2030 productivity rates22030Late scenarioMidpoint scenarioLate scenarioMidpoint scenario 19721982 20022012 201220221.41.22.00.91.0 19821992 199220020.20.22.84.10.10.10.81.91.13.00.10.142Generative AI and the future of work in AustraliaNevert

258、heless,as in many developed economies,Australia faces a slowdown in productivity growth.Even if it realized half of the productivity potential in the midpoint automation scenario,Australia would be back in line with the productivity growth of its post-1990s heyday.Gen AI offers a potential boost to

259、stagnating productivity levels,but like other technology before,it might not quickly achieve the direct translation of potential into saving time spent on activities.Organizations may need to apply concerted efforts to reap productivity benefits.As such,achieving midpoint adoption could be an optimi

260、stic scenario for Australia.In countries such as Finland and the United States,investment in AI as a proportion of GDP is about double that in Australiaand the relative AI skills penetration rate is higher in India(3.23 percent),Germany(1.72 percent),Singapore(1.37 percent),and Italy(0.97 percent)co

261、mpared to Australia(0.89 percent).Australia also has an opportunity to improve on reskillingamong OECD countries,Australia ranks among the bottom 2040 percent for improving the use of workers skills.Compared to their peers in Canada,England,and the United States,fewer Australians participate in form

262、al adult education and trainingand in a Business Council of Australia survey cited by the Treasury,72 percent of respondents reported that they had not received any training or professional development in the previous two years.This matters because successful organizational change may require teachi

263、ng employees new skills as well as gaining employee buy-in to use their new skills and tools.Ineffective reskilling efforts could widen existing occupation and skills gaps,and dilute the value proposition of the automation opportunity.In many parts of the economy,employers may need to make significa

264、nt efforts to engage employees and train them on how to make the most of AI in their newly transformed jobs.How gen AI could transform workforces,operations,and customer service in key sectors Rather than imagining an entire economy using automation and gen AI,it is more helpful to envisage the chan

265、ges to individual sectors,occupations,and businesses.To illustrate the potentially transformative impact of gen AI on operations,customer service,and productivity,we took a closer look at three key sectors of the Australian economy:the retail trade,financial services and insurance,and the public sec

266、tor(Exhibit 16).Alongside analysis of skill shifts,we present gen AI use cases drawn from interviews with sector representatives.32 Intergenerational report 2023Australias future to 2063,Commonwealth of Australia,August 24,2023.33 The AI Index 2023 Annual Report,AI Index Steering Committee,Institute

267、 for Human-Centered AI,Stanford University,April 2023.34 2019 OECD Skills Strategy:Australia,OECD,May 22,2019.35 Working Future:The Australian Governments White Paper on Jobs and Opportunities,“Chapter 5Filling skills needs and building our future workforce,”The Treasury,September 25,2023.Even if it

268、 realized half of the productivity potential in the midpoint automation scenario,Australia would be back in line with the productivity growth of its post-1990s heyday.43Generative AI and the future of work in AustraliaIn these sectors,approximately 30 percent of task hours could be automated by 2030

269、,with as much as one-third of that potential driven by gen AI.If that potential is met,the mix of necessary skills in each sector would change significantly.For example,the retail trade sector could see less need for tasks that use only basic cognitive skills such as data entry,and greater need for

270、social and emotional skills.Financial services could see a similar decline in the need for basic skills such as manual data entry,and a rise in demand for advanced communication and negotiation skills.And in the public sector,digital skills could become much more important,along with creativity,crit

271、ical thinking,and decision making.The result could be significant shifts in demand for particular jobs in these sectors.In financial services,for example,roles such as bank tellers,procurement clerks,and administrative support could see a marked decline in demand.But demand is likely to increase for

272、 business and financial specialists,managers,and computer engineers.Exhibit 16Shift in skills demands across financial,retail,and public sectors:Less basic cognitive,more social,emotional,and digital skills.Evolution of skill mix by sector 202230,share of total hours,%Evolution of 25 skill mixChange

273、 in hours,%RetailPublicFinancialRetailPublicBasic data input and processingBasic literacy,numeracy,and communicationAdvanced literacy and writingComplex information processing and interpretationCreativityCritical thinking and decision makingProject managementQuantitative and statistical skillsFine m

274、otor skillsGeneral equipment operation and navigationGeneral equipment repair and mechanical skillsGross motor skills and strengthCraft and technician skillsInspecting and monitoringAdvanced communication and negotiation skillsEntrepreneurship and initiative-takingInterpersonal skills and empathyLea

275、dership and managing othersAdaptability and continuous learningTeaching and training othersAdvanced IT skills and programmingBasic digital skillsScientifc research and developmentAdvanced data analysis and mathematical skillsTechnology design,engineering,and maintenanceFinancial202220302022203020222

276、0302243032226532275%+25%to 75%0%to 25%0%to 10%10%to 50%Basic cognitivePhysical and manualHigher cognitiveSocial and emotionalTechnological Note:Figures may not sum to 100%,because of rounding.Source:Australian Bureau of Statistics;O*NET;McKinsey Global Ins

277、titute analysis44Generative AI and the future of work in AustraliaThe effects of automation could intersect with other macro trends,such as the acceleration of e-commerce.For example,demand is estimated to decline for sales workers,who make up about half of all employees in Australias retail trade;w

278、e estimate that the sector could need 37,500 fewer sales roles by 2030.In contrast,demand for physical and manual workerssuch as material movers and loaderscould increase by approximately 30 percent in the sector.Retail trade:Opportunities to delight customers and deliver personalized serviceIn the

279、retail trade sector,automation and gen AI could enable employees to spend more of their time and energy on connecting with and delighting customers.Picture,for example,a retail store where shoppers converse with gen AI-powered information kiosks that advise on the most suitable products to match the

280、ir needsoffering price and feature comparisons.With automated checkout systems installed at the exits,shoppers pay for their goods by simply carrying their selected items out of the store.Such technologies could allow customer service representatives to spend more time advising customers on complex

281、queries,and less time behind the till or providing basic information.Retailers could also run their stores with fewer employees per shift,without compromising customer relationships.Gen AI could also facilitate a more personalized customer experience.Virtual assistants,for example,could tailor the e

282、nd-to-end shopping journey to the customers specific preferences and requirements,potentially increasing customer loyalty and sales.An example is Walmarts“text to shop”feature,which uses a chatbot to create a shopping experience that mirrors the familiar feeling of a text message conversationfor exa

283、mple,on which food to pick for a themed event or type of cuisine.Gen AI can also transform customer experience for pure-play e-commerce retailers that focus on only one industry.Adore Beauty,which sells more than 260 beauty brands online,is an example:it used Amazon Web Services(AWS)technology to bu

284、ild a gen AI application to summarize large quantities of product reviews,at speed and scale,into pros and cons for the benefit of both employees and customers.There are also opportunities to harness gen AI to transform customer experience in adjacent sectors.In telecommunications,for example,Telstr

285、a is exploring how gen AI can reduce wait times and provide a sense of continuity for customers.In a recent pilot,the company used OpenAI to distill customer transactions and interactions into one-sentence summaries,which employees can then refer to in future interactions with the same customer.The

286、result is that,in every interaction,frontline employees can understand and empathize with the customers experience with Telstra so far,quickly ascertain their needs,and take the right action.Financial services and insurance:Strengthening operations,reducing risk,and serving customers better In finan

287、cial services,gen AI could have a transformative impact across the banking and insurance value chains.One key opportunity lies in automating operations and customer service,in areas such as credit assessments and generating loan contracts.Gen AI could also augment employee performance by providing a

288、lways-on technical support through bots trained on policies,research,and customer 36 Lauren Forristal,“Walmart experiments with generative AI tools that can help you plan a party or decorate,”TechCrunch,October 4,2023.45Generative AI and the future of work in Australiainteractions,as well as bots th

289、at monitor clients financial needs.Financial services companies could also use gen AI to develop personalized marketing and sales content tailored to clients based on their profiles and history.In Australia,banks and insurance providers have been leaders in automation.The Commonwealth Bank of Austra

290、lia(CBA),for example,has announced initiatives that use gen AI across its business.These include Gen.ai studio,a tool used in call centers to suggest responses to operators in real time.CBA has also begun to develop chatbots to provide customized messaging to customers,while supporting employees wit

291、h product services and testing.Other banks and insurers have embraced gen AI too.For example,Macquarie has employed gen AI in its risk-management processes to enhance customer protection,and has integrated gen AI into call centers to answer customer queries.Public sector:Driving smarter decision mak

292、ing and more effective citizen services With many governments facing backlogs and long wait-lists for public services,gen AI could be a key lever to alleviate this pressureand for citizens,gen AI could make it much easier to navigate such services.As outlined in McKinseys article“Unlocking the poten

293、tial of generative AI:Three key questions for government agencies,”each government department or agency could see a range of transformative use cases.In tax,for example,gen AI could serve as a virtual assistant for real-time document verification,and a public assistant for real-time tax filing guida

294、nce.It could also generate risk assessment summaries from diverse data sources,prepare preliminary audit reports for potential areas of noncompliance,and prepare first drafts of personalized notifications to taxpayers and traders.In defense and intelligence,gen AI could act as a real-time translator

295、 for intelligence agencies,draft initial intelligence operation reports,tailor training content to specific military missions,simulate potential conflict scenarios for strategic planning,and generate synthetic misinformation data to enhance surveillance AI.(This being an area of low risk tolerance,h

296、uman analysts would need to remain closely involved in such activities.)In many parts of the world,AI integration is helping citizens seek out information and access government services with greater ease.For example,the city of Heidelberg in Germany has launched the Lumi chatbot,the countrys first d

297、igital citizen assistant.The tool enables people to easily navigate government services such as applying for a new identity card,getting a drivers license,and registering a place of residence.Adoption of these types of use cases may require governments to first build trust and confidence that the te

298、chnology can be implemented in a way that protects citizen privacy and is accurate and fair.In practice,back-office use cases may be deployed first,with citizen-facing use cases to follow as confidence builds.Across all of these use cases,governments could ensure they balance the adoption of gen AI

299、with proper safeguards,so that data is used appropriately and risks are minimized.We consider these issues further in the next chapter.37 Kate Weber,“CBA brings CommSec into its banking app,”iTnews,May 24,2023.38 Ry Crozier,“Macquarie Bank taps Google Cloud for AI work,”iTnews,June 21,2023;James Eye

300、rs,“Macquarie takes an AI-first approach in banking,”Financial Review,June 21,2023.39“Unlocking the potential of generative AI:Three key questions for government agencies,”McKinsey,December 7,2023.40“AI citizen assistance Lumi:The chatbot for all questions about Heidelberg,”City of Heidelberg,2024.4

301、6Generative AI and the future of work in AustraliaConsiderations for employers,governments,and educators 4.Automation driven by gen AI has the potential to transform workplaces across Australia.Employers can harness technology to improve productivity,create value,and help their employees navigate th

302、e change;governments can implement policies to unlock the benefits of automation;and educators can better meet the evolving needs of employers and use gen AI to reimagine how they teach.Employers can rethink strategy,talent,and capabilities to create value from automation Employers can balance speed

303、 of adoption with sustainability to maximize productivity gains while supporting employees in the transition.We suggest the following questions and key considerations.1.How could gen AI affect our competitive advantage and value proposition?Machines accelerated capabilities mean that all employers m

304、ay need to take a step back to consider what makes them special in a competitive and rapidly evolving market.For some organizations,the direct impact of gen AI may be readily apparent.Law firms,for example,can reconsider what makes them distinctive in a world where AI can generate first drafts of co

305、ntracts,automate many of the tailoring elements,and identify contentious,inconsistent,or missing clauses.Where businesses rely on cost competitiveness,they may use new technology to reduce their input costs.At a law firm,for example,the cost of contract development could shift from a service-driven

306、activity with variable billable hours to a product with a set fee.Where organizations rely on outstanding customer experience,the smart use of gen AI could enhance the personalization that customers have come to expect.2.Do we have a strategic workforce plan that matches demand and supply with the c

307、apabilities we need?Informed by clear thinking on their future value proposition and source of competitive advantage,organizations can identify their future skills needs and develop a plan to address gaps between the current state and their aspiration.Determine the skills required in the futureReali

308、zing new or redefined competitive advantages may require new skills.A creative agency,for instance,might need fewer manual design skills,but not all design decisions can be automated;humans play an important role in making sure that AI-generated content aligns with clients brand guidelines.Such an a

309、gency could require new skills such as the ability to work with gen AI algorithms to input prompts,interpret the outputs,and incorporate them into creative deliverables.Assess current capabilities and forecast future talent demand and supplyWith future skill requirements in focus,employers can condu

310、ct a skills audit to baseline current capabilitiesincluding both technical capabilities,and social and emotional skills.This assessment is a critical step in enabling organizations to identify skills gaps and work to close them.Organizations can forecast talent supply and demand in their respective

311、industries to anticipate which skills and roles might be most difficult to find,and they can begin formulating talent plans to address this.Rather than taking a“set and forget”approach,organizations could consider this assessment an ongoing exercise to be updated regularly as technologies and macro

312、trends evolve.48Generative AI and the future of work in AustraliaDevelop a strategy for attracting,retaining,and reskilling talentHaving mapped out their expected skill requirements,organizations can refine their talent strategy to match their needs.In a competitive market with shortages in critical

313、 technology roles,organizations can consider a holistic approach to managing talent.IT-related roles may require particular attention.McKinseys Organizational Health Index research has shown that IT workers tend to score below the average on measures of organizational health.Strategies to improve IT

314、 workers engagement include investing in high-quality planning and development tools to streamline engineers workflows and developing satisfying career pathways that lead to increasingly intricate technical challenges.Gen AI is an emerging field,and not every technology team member may be equipped w

315、ith the skill to capture its potential.To tackle this challenge,leaders may need to think creativelyfor example,they might invest in gen AI training programs for existing teams to bring them up to speed,or partner with companies that have the required expertise.Reskill employees to translate gen AI

316、potential into adoption By 2030,most workers can expect to have at least one daily task that is automated or augmentedfor 77 percent of workers in Australia,20 percent of their task hours could be automated.Gen AI has the potential to lift employees productivity,but leaders may need to build skills

317、internally to translate that potential into adoption.For many organizations,reskilling could be critical,given that many of the required skills may not be readily available for hire,and the need to retain employees with deep organizational knowledge and experience.Some organizations in Australia hav

318、e already moved quickly.For example,CBA has announced a reskilling program for employees interested in data science,with the aim of teaching frontline workers to work with complex data and coding languages.One approach to improve the effectiveness of gen AI is to ensure that the rollout of automatio

319、n technologies incorporates prompt-engineering trainingthe practice of providing inputs for gen AI tools that will produce optimal outputs.3.How can we plan to scale and embed the technologies from day one?To ensure that gen AI is adopted and scaled successfully in their organizations,leaders can ma

320、p out specific paths to capture the value that automation presentsand manage the associated risk.They can support this with clear communication to employees.41 For more on organizational health,see Sven Blumberg,Ranja Reda Kouba,Suman Thareja,and Anna Wiesinger,“Tech talent tectonics:Ten new realiti

321、es for finding,keeping,and developing talent,”McKinsey,April 14,2022.42 Ibid.43 MGI analysis.44“Tech jobs update May 2023,”TechCouncil of Australia,May 2023.45 Richelle Deveau,Sonia Joseph Griffin,and Steve Reis,“AI-powered marketing and sales reach new heights with gen AI,”McKinsey,May 11,2023.49Ge

322、nerative AI and the future of work in AustraliaCapture the intended value from gen AIand manage the risksGen AI has the potential to unlock significant value for organizations:a recent study,for example,found that software engineers completed coding tasks up to twice as fast when using gen AI and re

323、ported greater satisfaction when using the process.To capture that value,however,organizations may need to rethink business processes and redesign their IT infrastructure.Clear and consistent measurement of value creationand the investment needed to achieve itcan enable the right focus,accountabilit

324、y,and transparency.For example,organizations can closely monitor the ongoing running costs of gen AI,as the computing power required can create significant costs as the model is scaled.It could be just as important to manage the risks associated with gen AIfor example,organizations may need to reeva

325、luate their existing security protocols,understand how cybercriminals might use gen AI to carry out sophisticated cyberattacks,then take the necessary mitigating action.Use clear communication to engage employees on gen AI As the use of gen AI is scaled across the organization,clear communication ca

326、n become ever more important.Consistent messaging on the value of gen AI for both productivity and worker gains can alleviate fears that“the robots are taking over”and allow employees to focus on how technology can support their work(see sidebar“Using gen AI responsibly”).One effective communication

327、 approach is to involve employees in the initial decision on gen AI use cases.For instance,Lendi Group,a large Australian retail mortgage broker,uses a proprietary gen AI platform,supported by AWS technology.The company invited over 100 of its employees to suggest how they should use gen AI and to v

328、ote on which use cases they wanted to adopt first.David Hyman,Lendis CEO,said:“We think this could be a fundamental step change opportunity and for it to become that;the whole organization needs to think about generative AI and where it can deliver value for our business.”Another key step in employe

329、e communication is to create clear guidelines to mitigate the risk of infringing privacy and intellectual property,and of generating harmful,inaccurate,or biased information through gen AI.46“Unleashing developer productivity with generative AI,”McKinsey,June 27,2023.Consistent messaging on the valu

330、e of gen AI for both productivity and worker gains can alleviate fears that“the robots are taking over”and allow employees to focus on how technology can support their work.50Generative AI and the future of work in AustraliaUsing gen AI responsiblyThe challenge for employers,governments,and educator

331、s is balancing the benefits with the potential challenges and limitations of using gen AI technologies.Through thoughtful design and deployment,and maintaining continuous human oversight,organizations can prepare to use gen AI responsibly,safely,and ethically.The art is in crafting a robust approach

332、 to risk without being immobilized by caution,which could leave a lot of value on the table.One way of achieving this balance is involving senior risk stakeholders from day one.Enhance privacy.Confidential data may be leaked if users introduce such information into gen AI models through promptswhich

333、 could resurface in future outputs containing personally identifiable information.Organizations could mitigate this risk in part by being thoughtful about use case selection;removing any personally identifiable information in data preparation;and adopting data protection,privacy,and security guidelines and systems.If using public models,users can be supported by clear communication on what is and