《普华永道（PWC）：2024年全球人工智能就业晴雨表（英文版）（31页）.pdf》由会员分享，可在线阅读，更多相关《普华永道（PWC）：2024年全球人工智能就业晴雨表（英文版）（31页）.pdf（31页珍藏版）》请在三个皮匠报告上搜索。

1、How will AI affect jobs,skills,wages,and productivity? 2024 AI Jobs Barometer2PwCs 2024 AI Jobs BarometerSectors with highest AI penetration are seeing almost fivefold(4.8x)greater labour productivity growth.Rising labour productivity can generate economic growth,higher wages,and enhanced living sta

2、ndards.Jobs that require AI specialist skills carry up to a 25%wage premium in some markets.Skills sought by employers are changing at a 25%higher rate in occupations most able to use AI.To stay relevant,workers in these jobs will need to build or demonstrate new skills.Growth in jobs that require A

3、I specialist skills has outpaced all jobs since 2016(well before ChatGPT brought fresh attention to AI),with numbers of AI specialist jobs growing 3.5 times faster than all jobs.4.8x25%25%3.5xPwCs 2024 AI Jobs Barometer goes beyond predictions about AIs impact to find evidence by analysing over half

4、 a billion job ads.The Barometer reveals how AI is transforming the world of work,making people and businesses more productive while changing what it takes for workers to succeed.Headline Findings3PwCs 2024 AI Jobs BarometerAI is the Industrial Revolution of knowledge work,transforming how all worke

5、rs can apply information,create content,and deliver results at speed and scale.How is this affecting jobs?With the AI Jobs Barometer,PwC set out to find empirical evidence to help sort fact from fiction.PwC analysed over half a billion job ads from 15 countries to find evidence of AIs impact at worl

6、dwide scale through jobs and productivity data.PwC tracked the growth of jobs that demand specialist AI skills(such as machine learning or neural networks)across countries and sectors as an indication of AI penetration.1 We find that AI penetration is accelerating,especially in professional services

7、,information&communication,and financial services.Workers with specialist AI skills command significant wage premiums,suggesting that their abilities to deploy AI are valuable to companies.Half a billion job ads revealAIs impact1 AIs true penetration into the economy may be even greater than reflect

8、ed in this analysis.By focusing on job ads,this analysis captures AIs impact on job changers,but does not capture AI usage or upskilling for existing employees.4PwCs 2024 AI Jobs BarometerBut AIs impact is not limited to only those workers who have specialist AI skills.Many,if not most,workers who u

9、se AI tools in their work do not have or need these specialist skills.For example,a limited number of workers with specialist AI skills may design an AI system or tool for a company that is then used by hundreds or thousands of the companys customer service agents,analysts,or lawyers-none of whom ha

10、ve specialist AI skills.In fact,one thing that makes a well-known form of AI-generative AI-such a powerful technology is that typically it can be operated using simple everyday language with no technical skills required.To capture AIs impact on all jobs,PwC analysed all jobs(and sectors)by their lev

11、el of AI exposure.A higher level of AI exposure means that AI can more readily be used for some tasks.Examples of occupations with higher AI exposure are financial analysts,customer service agents,software coders,and administration managers.The analysis revealed that sectors with higher AI exposure

12、are experiencing much higher labour productivity growth.At the same time,the skills demanded by employers in AI-exposed occupations are changing fast.Read on to learn more.The AI Jobs Barometer uses half a billion job ads from 15 countries to examine AIs impact on jobs,skills,wages,and productivity5

13、PwCs 2024 AI Jobs BarometerKey TermsAI specialist skills:Specialist,technical AI skills like deep learning or cognitive automation.See Appendix One for AI skills list.AI specialist jobs:Jobs that require specialist,technical AI skills.All jobs:All jobs in all occupations.AI-exposed:Describes all job

14、s or sectors in which AI can readily be used for some tasks(based on definition of AI Occupational Exposure developed by Felten et al.)6PwCs 2024 AI Jobs BarometerAttention to AIs impact on the jobs market exploded in November 2022 with the launch of ChatGPT 3.5.However,the data shows that AI had qu

15、ietly exerted a growing impact on the jobs market years before.Growth in AI specialist jobs has outpaced growth in all jobs since 2016,well before ChatGPT brought fresh focus to AI.Today,there are seven times as many postings for specialist AI jobs as there were in 2012.In contrast,postings for all

16、jobs have grown more slowly,doubling since 2012.Put another way,openings for jobs that require specialist AI skills have grown 3.5 times faster than openings for all jobs since 2012.The 2022 peak in job postings above represents exceptionally high demand for workers which gradually eased in 2023 as

17、job market conditions returned toward normal.Growth in Al jobs has outpaced all jobs since at least 2016Source:PwC analysis of Lightcast data.The analysis represents six of the fifteen countries:US,UK,Singapore,Australia,Canada,and New Zealand.Nine countries have been excluded due to data prior to 2

18、018 being unavailable:France,Germany,Belgium,Denmark,Spain,Italy,Netherlands,Norway,and Sweden.Number of Job Postings,relative to 2012 AI Jobs All Jobs2020000222023AI penetration isaccelerating7PwCs 2024 AI Jobs BarometerKnowledge work sectors in particula

19、r are seeing growing demand for jobs that require specialist AI skills.The share of job ads requiring these skills is higher in professional services,information&communication,and financial services-precisely those sectors predicted to be most exposed to AI.2 Financial services has a 2.8x higher sha

20、re of jobs requiring AI skills vs other sectors,professional services is 3x higher,and information&communication is 5x higher.2 AI Occupational Exposure(AIOE),constructed by Felten et al(2021),scores and measures the degree to which occupations rely on abilities in which AI has made the most progres

21、s in recent years,meaning AI can more readily be used for some tasks in those occupations.Share of job postings by sector requiring Al skillsSources:PwC analysis of Lightcast data,UK Government Impact of AI on Jobs 2023.“Other sectors”includes Agriculture,Mining,Power,Water,Retail trade,Transportati

22、on,Accommodation,Real Estate,Administrative activities,Arts and Entertainment,Household activities,Construction,Manufacturing,Education,Health and Social Activities and Extraterritorial Activities sectors.Chart includes all 15 countries in this study.5%4%3%2%1%Information and Communication Professio

23、nal Services Financial Services Other sectors20120%20000222023Knowledge work sectors have higher AI penetration8PwCs 2024 AI Jobs BarometerAIs value to companies is made clear by what is happening with the wages of workers with AI specialist skills-the very people wh

24、o are making the AI revolution possible.AI specialist jobs command up to a 25%wage premium on average25%Up to 25%wage premium for workers with specialist AI skills9PwCs 2024 AI Jobs BarometerWage premium for job vacancies which require Al skills by countrySources:PwC analysis of Lightcast data,ISCO-

25、08 Occupation Codes(4-digit level).2023 data.These findings do not demonstrate a causal relationship.These estimates are calculated by comparing the average salaries of AI job postings to those of non-AI postings for the same occupations.Two filters are applied to ensure(1)the count of AI job postin

26、gs and(2)the ratio of AI jobs:non-AI jobs being compared is above a certain threshold.The analysis provided represents five of the 15 countries:UK,USA,Singapore,Canada and Australia.The remainder of the countries have been omitted from this analysis as the data was less extensive:New Zealand,Italy,F

27、rance,Germany,Spain,Belgium,Netherlands,Denmark,Norway and Sweden.Country Al Wage PremiumOccupationUSAUKCanadaAustraliaSingaporeDatabase Designers and Administrators+53%+58%+8%+14%+35%Lawyers+49%+27%-5%Sales and Marketing Managers+43%+14%+3%+7%+3%Financial Analysts+33%+32%-+11%Applications Programme

28、rs+32%+24%-+7%+34%Systems Analysts+30%+34%+15%+7%+28%Accountants+18%+5%+17%-+4%Average wage premium across all jobs+25%+14%+11%+6%+7%For example,job ads for US sales managers that require AI specialist skills offer wages that are on average 43%higher than job ads for sales managers that do not requi

29、re AI skills.Canadas accountants can enjoy a 17%wage premium if they have AI specialist skills,and UK employers are willing to pay a 27%premium for lawyers equipped with AI skills.As we have seen,growth in jobs demanding AI specialist skills has outpaced growth in all jobs since 2016.Whats more,thes

30、e jobs carry up to a 25%wage premium on average,underlining the value of these skills to companies.Below are average AI wage premiums for five countries for which there is sufficient data to perform the analysis.To show how this wage premium can affect individual occupations,wage premiums for select

31、ed occupations are given.10PwCs 2024 AI Jobs BarometerSo far this report has discussed jobs which require specialist AI skills like deep learning or natural language processing.But many,if not most,workers who use AI tools in their work do not have these skills.To understand how AI is affecting all

32、jobs,PwC examined jobs and sectors by their levels of AI exposure which means the degree to which AI can readily be used for some tasks.PwCs analysis revealed how higher levels of AI exposure appear to be affecting workers productivity,numbers of job openings,and the skills that jobs require.First,l

33、ets see how AI may be affecting productivity.Labour productivity growth has been sluggish in many nations for years.OECD countries have experienced a lost decade of labour productivity growth with weak average annual rises of 1.1%from 2011 to 2020,followed by declines in both 2021 and 2022.3 3 OECD,

34、Labour Productivity and Utilisation.The pandemic had a negative impact on productivity in 2020-2022.AI appears to be driving aproductivity revolution4.8xhigher growth in labour productivity in Al-exposed sectors11PwCs 2024 AI Jobs BarometerAl exposure and labour productivity growth rate by sector.Ea

35、ch dot represents a country.Sources:PwC analysis of OECD data,Felten et al.(2021).The AI Occupation Exposure(AIOE)constructed by Felten et als(2021)AI Occupational Exposure(AIOE)scores and measures the degree to which occupations rely on abilities in which AI has made the most progress in recent yea

36、rs,meaning AI can more readily be used for some tasks.The AIOE score is a relative measure,where higher numbers indicate greater exposure to AI,meaning that even negative values still imply a certain degree of exposure to AI.To measure the growth rate in labour productivity,PwC used the OECDs GVA pe

37、r person employed metric,indexed on 2018.Due to the availability of the OECD data,PwC focused on just six sectors.The 2023 OECD labour productivity data has not been released.Therefore the labour productivity growth rate between 2018 and 2022 is considered.If the view that AI is increasing productiv

38、ity is correct,it would be expected that the pattern of stronger productivity growth for AI-exposed sectors would continue or accelerate in 2023.The 4.8x higher growth is a comparison of averaged labour productivity growth rates;absolute growth rates are 0.9%and 4.3%.Labour productivity growth rate

39、%20%10%0%-10%-20%-1.5-1-0.500.511.522.5Sectors less exposed to AlAl Sectoral ExposureSectors more exposed to AlAverage.9%Average 4.3%ConstructionFinancial ServicesProfessional servicesInformation TechnologyManufacturingRetail,Food Services&TransportThis stagnant labour productivity is a s

40、erious problem because it is a drag on economic growth,reducing potential tax revenues,chipping away at investment in public services and flatlining living standards.Recently there has been much speculation that AI can supercharge workers productivity.The good news is there is now evidence to sugges

41、t that this is not just wishful thinking,and is already fast becoming reality.We have seen that three sectors-financial services,IT,and professional services-have higher AI exposure and higher AI penetration.How is this affecting productivity?The data shows that these three sectors are seeing nearly

42、 5x faster productivity growth than sectors with lower AI exposure(such as transport,manufacturing and construction).12PwCs 2024 AI Jobs BarometerWhile it is not possible to prove causation,this is an intriguing pattern.Unlike the computer revolution which took significant time to enhance productivi

43、ty(economist Robert Solow once observed that the impact of the computer age was evident everywhere but in the productivity statistics),the data suggests AI is already doing so,right now.AI may be compressing the productivity J-curve4 in which new technologies can take significant time to cause a sha

44、rp uptick in productivity.PwCs 2024 Global CEO Survey confirms that 84%of CEOs whose companies have begun to adopt AI believe it will increase efficiency in their employees time at work.5 Increasing productivity means more than just doing the old things faster.It also means finding new,AI-powered wa

45、ys to create value.In fact,70%of CEOs say that AI will significantly change the way their company creates,delivers and captures value over the next three years.The implications for business are huge.Global CEOs anticipate that one form of AI-generative AI-will deliver significant top and bottom line

46、 benefits,with 46%saying it will increase profitability,and 41%saying it will increase revenue.Investors agree.PwCs 2023 Global Investor Survey shows that investors believe accelerated adoption of AI is critical to the value equation,with 61%of investors saying faster adoption is very or extremely i

47、mportant.When responses indicating moderately important are included,the proportion jumps to 85%.All of this adds up to a positive story for the global economy:a revolution in productivity and value creation.4 Productivity J-curve,Brynjolfsson et al.,National Bureau of Economic Research.5 Around a t

48、hird of the respondents in our 2024 Global CEO Survey have begun to adopt AI.Of these,84%believe it will increase employees efficiency.These findings suggest that companies leading the way on AI deployment are seeing the benefits.Al does more than help workers do the old things faster.Al opens the d

49、oor to new business models and ways of creating value.13PwCs 2024 AI Jobs BarometerIn AI-exposed occupations such as customer services and IT-a number of which have acute labour shortages-jobs are still growing,but 27%more slowly on average.This could be good news for many nations facing shrinking w

50、orking age populations and vast unmet needs for labour in many sectors.AI can help to ensure that the labour supply is available for the economy to reach its full potential.27%Lower job growth in Al-exposed occupations(though jobs still growing overall)AI is helping to easelabour shortages14PwCs 202

51、4 AI Jobs BarometerIt is important to emphasise that job numbers in AI-exposed occupations are still growing.The data suggests that AI does not herald an era of job losses but rather more gradual jobs growth,helping to enable companies to find the workers they need.Job openings are still growing in

52、Al-exposed occupations,but more slowlySources:PwC analysis of Lightcast data,ISCO-08 Occupation Codes(2-digit level)and Felten et.al AI Occupation Exposure.The cross-country comparison on the right hand side considers the difference in the growth in job postings between the occupations most exposed

53、to AI and those least exposed to AI.Growth rate in postings %200%150%100%50%0%-50%-100%-2.00-1.50-1.00-0.500.000.501.001.502.00Occupation less exposed to Al Average growth 80%Occupation is more exposed to Al Average growth 58%Al Occupation ExposureSales and Service WorkersBusiness Profes

54、sionalsIT ProfessionalsAdministrative&Commercial ManagersPlant and Machinery OperatorsConstruction and Manufacturing LabourersClerical Support WorkersAgriculture LabourersCleaners and Helpers15PwCs 2024 AI Jobs BarometerThe skills required by employers in AI-exposed occupations are changing fast.Old

55、 skills are disappearing from job ads-and new skills are appearing-25%faster than in roles less exposed to AI.25%higher skills change in Al-exposed occupationsWhat this means for workers:Build skills for an AI age16PwCs 2024 AI Jobs BarometerWorkers in AI-exposed roles may need to demonstrate or acq

56、uire new skills to stay relevant in a jobs market that is fast-evolving.PwCs 2024 Global CEO Survey makes it clear that 69%of CEOs anticipate that generative AI will require most of their workforce to develop new skills,rising to 87%of CEOs who have already deployed generative AI.Workers need to tak

57、e ownership of their learning,rapidly developing the skills to remain relevant and to embrace the opportunity AI brings.Change in skills demanded by employers for occupations more(and less)exposed to AlSources:PwC analysis of Lightcast data,ISCO-08 Occupation Codes(2-digit level),Felten et al.(2021)

58、.The net skill change is based on Deming and Noray(2020)and is calculated by using the difference between 2019-2023 in the total number of skills required by job occupations using the ISCO-08 4-digit occupational codes.The AI Occupation Exposure is from Felten et als(2021)and measures the degree to

59、which occupations rely on abilities in which AI has made the most progress in recent years,meaning AI can more readily be used for some tasks.The correlation coefficient is.31 and is the statistical measure that quantifies the strength and direction of a linear relationship between net skill change

60、and AI Exposure.To calculate the average net skill changes for the most and least exposed occupations to AI,an average of the net skill change of the top and bottom quartile of occupations is taken based on their exposure to AI.See Appendix Two for formula.Net skill change between 012

61、14-2-1.5-1-0.500.511.52Occupation less exposed to Al Average:2.7(Least exposed quartile)Greater change in skills demanded by employerSmaller change in skills demanded by employerOccupation is more exposed to Al Average:3.4(Most exposed quartile)Al Occupation ExposureWeb and multimedia developersSoft

62、ware developersAthletes and sports playersHand launderersRoofersMathematicians,actuaries and statisticiansJudgesPsychologistsWeb technicians17PwCs 2024 AI Jobs BarometerThere are clues to which skills workers may want to build to prosper in an AI age.Some of the skills rising fastest in demand are t

63、hose which cannot easily be performed by AI.Below are four of the skills categories rising fastest in demand,and for each category a few examples are provided of specific skills with growing demand.From dam construction to sports instruction,some skills with booming demand are relatively hard for AI

64、 to perform.On the other hand,what skills are declining in demand?Below are four skills categories with the steepest declines in employer demand,with a few illustrative examples of specific skills with falling(or rising)demand within each category.The AI transformation is clear to see in categories

65、like Information Technology where demand for AI-related skills like AI/Machine Learning Inference6 is flourishing,while FASTEST growing skill categoriesSources:PwC analysis of Lightcast data.Data based on 2019-2023.The overall growth in skill categories is calculated as the change in the average sha

66、re of the skill category for all countries between 2019 and 2023.YogaSport InstructorsSwimmingCreative ArtsPerforming Arts,Sports,and RecreationPersonal Care and ServicesEnergy and UtilitiesEnvironmentSkill sub-categoryGrowth in skills sub-categoryChild SafeguardingLaser Hair RemovalSkin TreatmentsF

67、uneral ArrangementsSolar DevelopmentWater MeteringEnergy TradingDam ConstructionSediment SamplingEcological RestorationWaste CollectionFlood Controls+426%+178%+20%+18%+156%+84%+41%+11%+87%+58%+44%+33%+84%+57%+32%+17%+82%+155%6 AI/Machine Learning Inference means applying a machine learning model to

68、a dataset to generate an output,insight,or prediction.+48%+58%demand for some skills that may be more readily replaced by AI(such as coding in Javascript)is falling.The Analysis category shows a similar pattern with soaring demand for Natural Language Programming(an AI skill)and declining demand for

69、 Regression Analysis,a type of analysis AI can help to perform.SLOWEST growing skill categoriesSources:PwC analysis of Lightcast data.Data based on 2019-2023.The overall growth in skill categories is calculated as the change in the average share of the skill category for all countries between 2019 a

70、nd 2023.AI/ML InferenceSmart DevicesCloud OperationsJavascriptInformation TechnologyDesignSalesAnalysisSkill sub-categoryWithin the slowest growing skills categories,some sub-categories buck the trend and are growing fast.Some of these (like Al/ML Inference)are Al skills.Growth in skills sub-categor

71、yGame DesignVisual EffectsComputer GraphicsInterface DesignPipeline ManagementConsumer SalesOnline AuctionsCold CallingNatural Language ProgrammingAsset AnalyticsData SynthesisRegression Analysis+113%+81%-7%-37%+12%-11%-30%-46%+6%-11%-23%-37%+64%+3%-8%-21%-14%-20%-23%-26%19PwCs 2024 AI Jobs Baromete

72、rAI is redefining what it means to be a financial analyst,a software coder,a customer service agent(and many more roles),opening up whole new possibilities for workers to deliver impact.Workers who learn to harness AI are likely to have bright futures in which they can generate greater value and cou

73、ld consequently have greater bargaining power for wages-all within a context of rising societal prosperity.Workers agree.PwCs 2023 Global Workforce Hopes and Fears Survey shows workers expect mostly positive benefits from AI with 31%expecting AI to increase their productivity/efficiency and 21%expec

74、ting AI to create new job opportunities.Many who predict AI will cause a sharp decline in job numbers are asking the wrong question.Those who predict AI will have a negative impact on total job numbers often look backward,asking whether AI can perform some tasks in the same way as they have been don

75、e in the past.The answer is yes.But the right question to ask is this:How will AI give us the power to do entirely new things,generating new roles and even new industries?No going back to yesterdays jobs market-but vast opportunities await those whoadapt to an AI age20PwCs 2024 AI Jobs BarometerAI o

76、ften performs best in partnership with people.Without oversight,AI can miss context and nuance or give poorer quality output.Only humans can fully appreciate and navigate the people,processes,and context of individual organisations and situations.Our analysis(particularly the finding about AIs poten

77、tial impact on productivity)suggests that AIs effect on jobs may be similar to that of the internal combustion engine in the 20th century which reduced numbers of some jobs(such as horse trader)while at the same time creating far more jobs than it displaced(from truck driver to road engineer to traf

78、fic police).As technology gets better at being technology,humans can get better at being humans.There is clear evidence that AI often delivers the best outcomes when used in partnership with people.The AI era requires a new style of leadership,an openness to bold transformation and inventive thinkin

79、g about how AI and people together can create new forms of value.Carol Stubbings,Global Markets and TLS Leader,PwC UKAI makes human labour more relevant and valuable,opening up new opportunities for people to develop new skills and enter new roles.AI will create new jobs for people that we havent ye

80、t begun to imagine.Many of the fastest-growing jobs of today-from cloud engineer to digital interface designer-didnt exist 10 or 20 years ago and have been generated by technology.Like a spreadsheet or a saw,AI is a tool that makes people more powerful and capable.Workers who build the skills to har

81、ness AI will be more valuable than ever.Pete Brown,Global Workforce Leader,PwC UK21PwCs 2024 AI Jobs Barometer7 The Second Machine Age,Eric BrynjolfssonFar from heralding the end of jobs,AI signals the start of a new era in which workers can be more productive and valuable than ever.Instead of focus

82、ing only on how AI can take on some tasks formerly done by people,we should think inventively about how to make the most of AI to create new industries and new roles for people.Embracing AI in this way is one way to bring about continued positive outcomes for workers.Economist Eric Brynjolfsson obse

83、rved,If AI is used mainly to mimic humans,to replace humans with machines,it is likely to lead to lower wages and more concentration of wealth.If we use AI mainly to augment our skills,to do new things,then it is likely to lead to widely shared prosperity and higher wages.7AI provides much more than

84、 efficiency gains.AI offers fundamentally new ways of creating value.In our work with clients,we see companies are using AI to amplify the value their people can deliver.We dont have enough software developers,doctors,or scientists to deliver all the code,healthcare,and scientific breakthroughs the

85、world needs.There is a nearly limitless demand for many things if we can improve our ability to deliver them.Scott Likens,Global AI and Innovation Technology Leader,PwC US22PwCs 2024 AI Jobs BarometerThere is no going back to yesterdays jobs market,but-if carefully managed-the AI revolution could br

86、ing a bright future for workers and companies.Below are steps that companies,workers,and policymakers can take to help realise AIs promise to grow productivity and fuel rising shared prosperity.Here is what companies can do.Business leaders can embrace,experiment,and create new uses of AI.They can t

87、hink beyond using AI to do things the way they have been done in the past and instead use AI to generate new ways to create value.While AI can help to make existing processes more efficient,companies can realise even more benefit from AI by using it to reinvent business models or pioneer new product

88、 lines.Thinking inventively about how to use AI helps the company to be the disruptor rather than the disrupted,and it helps to create new opportunities for people.Next steps for companies,workers,policymakers23PwCs 2024 AI Jobs BarometerBusiness leaders should view AI as a complement to people that

89、 is best used with human oversight.Leaders should track the ever-shifting jagged frontier8 which marks where AI performs brilliantly versus where AI lacks capabilities or works best with human assistance.Companies can support employees to make the most of AI by offering training and helping them see

90、 how AI empowers them(and can even make their jobs more enjoyable by freeing them to work more autonomously and be more confident in their roles)9.Firms can consider hiring on the basis of candidates skills rather than focusing solely on their degrees,job history,or previous job titles.This helps fi

91、rms find the workers they need,and it helps workers more readily adapt to a fast-changing jobs market.A study by PwC and the World Economic Forum conducted across 18 economies shows that a skills first approach has the potential to expand the talent pool by 100 million people.Companies can take a sk

92、ills first approach for existing employees too,treating workers as people with sets of skills and talents that can be fluidly applied across the organisation.10 These skills based organisations can more flexibly deploy workers,helping both companies and workers adapt to the AI transformation while o

93、pening up broader talent pools,developing more resilient talent pipelines for the jobs of tomorrow,and achieving enhanced levels of employee motivation,satisfaction,performance,and retention.11,12 Workers,for their part,should embrace AI,experimenting with it and seeking ways it can complement and e

94、nable them in their work.13 Workers should build the skills to be sought after in an AI age(for example,skills that either complement AI or are hard for AI to do).Some workers may need to adapt more than others to succeed in an AI era;for example,some workers may need only a little training to adopt

95、 AI tools while other workers may need to move to new occupations which require more extensive retraining or upskilling.Workers,companies,and policymakers share responsibility for helping all workers adapt to an AI era.8 Navigating the Jagged Technological Frontier:Field Experimental Evidence of the

96、 Effects of AI on Knowledge Worker Productivity and Quality,Fabrizio DellAcqua et al,Harvard Business School working paper9 MIT Sloan Management Review:Achieving Individual-and Organisational-value with AI,2022.10 Skills based organisations are integrating skills throughout the talent management lif

97、ecycle by implementing skills-based training programs for upskilling and reskilling,as well as establishing skills-based career pathways for redeployment.11 Skills-based sourcing&hiring playbook,Rework America Alliance,2022 12 AI can help with skills based hiring by,for example,automatically generat

98、ing and updating skills profiles and working out adjacent skills people are likely to have or could readily learn.13 Workers whose companies do not offer AI tools can experiment with public AI tools like ChatGPT.Workers should not use proprietary company data on public tools,but public tools still p

99、rovide a wealth of opportunities to get to know AIs power.24PwCs 2024 AI Jobs BarometerSources:The IMFs AI Preparedness Index ranks countries preparedness to adopt AI based on four pillars:Digital Infrastructure,Innovation&Integration,Human Capital&Policies and Regulation&Ethics.Proportion of total

100、job vacancies requiring Al related skills by country,2012-20232.5%2.0%1.5%1.0%0.5%Singapore Denmark United States New Zealand Germany United Kingdom France All countries20120.0%20000222023Singapore has the highest proportion of Al related job vacancies increasing to

101、4.8%in 2023Policymakers can support workers with training/retraining and safety nets,and shape the education system to help prepare workers for an AI age in which critical thinking,creativity,and adaptability are likely to be key skills.Finally,policymakers can strive to make sure that growing prosp

102、erity from AI adoption is widely shared.Policymakers can encourage the use of AI to grow productivity and prosperity,for example by building the supportive policy environment,digital infrastructure,and skilled workforce to help realise AIs potential.Countries with the strongest growth in jobs that d

103、emand AI skills(an indicator of AI usage and penetration)offer lessons for policymakers in how to create an environment conducive to making the most of AI.The three countries in this study with the highest proportion of jobs that require AI skills are Singapore,Denmark,and the US.These are the same

104、three countries that top the IMFs AI Preparedness Index ranking which measures areas such as digital infrastructure,human-capital and labor-market policies,innovation and economic integration,and regulation and ethics.Policymakers who would like their people to benefit from the AI revolution should

105、take note.25PwCs 2024 AI Jobs BarometerKey areas for actionPolicymakers Encourage the use of Al to grow productivity and prosperity Ensure growing prosperity from Al adoption is widely shared Support the use of Al to augment rather than replace workers Support workers with training/retraining,worker

106、 protections,and safety nets Shape the education system to help prepare workers for an Al age Ensure the responsible use of Al with PwCs Responsible Al frameworkBusinesses Embrace,experiment,and pioneer new uses of Al.Think beyond using Al to do existing tasks.Instead,pioneer new ways to create valu

107、e.Build a climate of confidence and trust in using Al so people are more likely to welcome trying it.Use Al to complement people.Figure out where Al is best used in partnership with people.Hire on the basis of skills rather than only degrees or work history.Upskill workers to make the most of Al Use

108、 Al responsibly with PwCs Responsible Al frameworkWorkers Embrace and experiment.Seek ways to make the most of Al to complement and empower you Build skills to be sought after in an Al-driven jobs market(e.g.skills that complement Al or are difficult for Al to do)26PwCs 2024 AI Jobs BarometerPwCs an

109、alysis of over half a billion job ads worldwide suggests that AI may help make workers more productive,opening the door to rising prosperity for workers and nations.Like past technological revolutions from electricity to internal combustion engines,AI is changing what it takes for workers to succeed

110、-and those who adapt may enjoy vast new opportunities.In conclusion27PwCs 2024 AI Jobs BarometerAcknowledgementsPartner SponsorsCarol Stubbings Global Markets and TLS LeaderScott Likens Global AI and Innovation Technology LeaderPeter Brown Global Workforce LeaderContributorsBarret Kupelian Director,

111、Chief Economist Sarah Brown Director,Global Corporate Affairs Mehdi Sahneh Senior Manager,Economist Simon Oates UK Economics LeaderJustine Brown Director,Global Workforce Adam Deasy Senior Associate,Economist Nabil Taleb Senior Associate,Economist Harry Ingham Associate,Economist Dr.Ilhan Guner Acad

112、emic Advisor University of KentAdvisorsTom Pagram Partner,Artificial Intelligence Leader&Chief Technology Officer,PwC AustraliaJohan Jegerajan Partner,CEMEA and UK Consulting CTO,PwC UKDr.Alexis Crowe Lead,Geopolitical Investing practice,PwC USAshootosh Chand Partner,Digital&Emerging Technologies,Pw

113、C indiaEugnie Krijnsen Partner,Global Financial Services Advisory Leader,PwC NetherlandsIlana Golbin Blumenfeld Director,Emerging Technologies&Responsible AI Lead,PwC USMir Kashifuddin Partner,Data Risk&Privacy Practice Leader,PwC USTom Lewis Partner,Head of Commercial Technology,PwC UKMitra Best Pa

114、rtner,Technology Impact Leader,PwC USEuan Cameron Partner,UK Artificial Intelligence and Drones Leader,PwC UKMaria Axente UK Responsible AI and AI for Good Lead,PwC UKRob McCargow Director,UK Technology Impact Leader,PwC UKCalen Byers Partner,Financial Services-Asset&Wealth Management-Real Assets,Pw

115、C USParul Munshi Partner,Workforce Transformation PwC South East Asia Consulting,PwC SingaporePrasun Shah Partner,UK Tax Workforce,PwC UKJulia Lamm Partner,Workforce Transformation,PwC USAnthony Bruce Partner,Chair of Health Industries,PwC UKEyhab Abdeen Partner,Middle East Workforce,PwC Middle East

116、Bastiaan Starink Partner,Workforce,PwC NetherlandsPaul Kett Senior Adviser and Global Director Education and Skills,PwC UKPatrick Pugh Partner,Principal,Global Microsoft Alliance and Transformation Leader,PwC USBhushan Sethi Partner,Strategy&,PwC US Barbara Baarsma Chief Economist,PwC EuropeAmy Cai

117、Managing Partner,ESG,PwC ChinaJennifer Kosar Trust and Transparency Solutions Leader,PwC US28PwCs 2024 AI Jobs BarometerAppendix one:AI skills listAI jobs are defined here as jobs that require one or more technical AI skills.Here is the complete list of AI skills used:mlpack(C+Library)Nvidia JetsonW

118、atson ConversationRobotic SystemsIPSoft AmeliaApache MADlibLoss FunctionsDask(Software)PydataAdvanced RoboticsSeq2SeqWatson StudioVowpal WabbitMatrix FactorizationKaldiGoogle Cloud ML EngineSemi-Supervised LearningDlib(C+Library)Robotic Liquid Handling SystemsLexalyticsAmazon ComprehendSemantic Pars

119、ingAutomated Machine LearningDialogflow(Google Service)Word EmbeddingAWS SageMakerAmazon TextractNatural Language GenerationMachine Learning MethodsVoice User InterfaceTest DatasetsCyber-Physical SystemsPart-of-Speech TaggingImage SegmentationConvolutional Neural NetworksDeep Learning MethodsFeature

120、 LearningAutonomous VehiclesAzure Machine LearningRelationship ExtractionWord-Sense DisambiguationCognitive AutomationProgrammatic Media BuyingTransfer LearningLong Short-Term Memory(LSTM)Amazon LexCaffe2Conversational AIKernel MethodsAdversarial Machine LearningUnmanned Aerial Systems(UAS)Amazon Po

121、llyLanguage IdentificationGuidance Navigation And Control SystemsNatural Language UnderstandingFast.aiEnsemble MethodsTraining DatasetsMeta LearningSpeech SynthesisAutoencodersPose EstimationIntelligent Virtual AssistantVoice Assistant TechnologyGradient BoostingApache SINGAOracle Autonomous Databas

122、eMicrosoft LUISApache MXNetOpen Neural Network Exchange(ONNX)Azure Cognitive ServicesCognitive ComputingBot FrameworkTorch(Machine Learning)Gesture Recognition3D ReconstructionAutonomous Cruise Control SystemsAdvanced Driver Assistance SystemsMulti-Agent SystemsArtificial IntelligenceApplications Of

123、 Artificial IntelligenceArtificial Intelligence Markup Language(AIML)Amazon AlexaANTLROptical Character Recognition(OCR)Artificial General IntelligenceArtificial Intelligence SystemsArtificial Neural NetworksComputer VisionAssociation Rule LearningAutonomic ComputingAutonomous SystemNaive Bayes Clas

124、sifierClassification And Regression Tree(CART)CHi-Squared Automatic Interaction Detection(CHAID)Decision Tree LearningCluster AnalysisComputational IntelligenceComputational LinguisticsImage AnalysisDecision ModelsDialog SystemsDigital Image ProcessingDimensionality ReductionExpectation Maximization

125、 AlgorithmEmbedded IntelligenceEvolutionary Acquisition Of Neural TopologiesEvolutionary ProgrammingExpert SystemsFuzzy LogicGaussian ProcessGenetic AlgorithmGeneral-Purpose Computing On Graphics Processing UnitsHyperparameter OptimizationHandwriting RecognitionHidden Markov ModelContextual Image Cl

126、assificationImage MatchingInference EngineInformation ExtractionIntelligent AgentIntelligent ControlCognitive RoboticsIntelligent SystemsInteractive KioskKnowledge EngineeringKnowledge-Based Systems29PwCs 2024 AI Jobs BarometerLanguage ModelMachine Learning AlgorithmsLIBSVMMachine LearningReasoning

127、SystemsMachine TranslationMachine VisionMarkov ChainMotion AnalysisMotion PlanningNatural Language ProcessingNatural Language ProgrammingNatural Language ToolkitsNatural Language User InterfaceNearest Neighbour AlgorithmNamed Entity RecognitionNatural Language Processing SystemsNLTK(NLP Analysis)Omn

128、iPageOpenCVApache OpenNLPSorting AlgorithmKnowledge-Based ConfigurationQuestion AnsweringRecommender SystemsEye TrackingTensorFlowObject RecognitionVoice TechnologyPath FindingRemote SensingRobot FrameworkRobot Operating SystemsRobotic Automation SoftwareScreen ReaderSemantic SearchSentiment Analysi

129、sServomotorSemantic Interpretation For Speech RecognitionSoft ComputingSpeech EnhancementSpeech ProcessingSpeech Recognition SoftwareSpeech TechnologyStatistical Language AcquisitionSupervised LearningSupport Vector MachineSwarm IntelligenceText MiningTokenizationFeature SelectionImage SensorSpeech

130、RecognitionWekaReinforcement LearningShogunRobotic ProgrammingSirikitPath AnalysisSemantic AnalysisSLAM Algorithms(Simultaneous Localization And Mapping)Data ClassificationFacial RecognitionFeature EngineeringChatbotCollaborative FilteringVoice InteractionPredictionioN GramRandom Forest AlgorithmApa

131、che SparkText ClassificationRealsenseCaffeDeep LearningFace DetectionLatent Dirichlet AllocationAdaBoost(Adaptive Boosting)Theano(Software)Keras(Neural Network Library)CortanaDisambiguationDeeplearning4jChainer(Deep Learning Framework)Scikit-Learn(Python Package)PerceptronPybrainWord2Vec ModelsXgboo

132、stMnistObjective FunctionCudnnMicrosoft Cognitive Toolkit(CNTK)Recurrent Neural Network(RNN)BoostingBaiduGame AiDbscanFeature ExtractionImagenetApache MahoutLemmatizationLight Detection And Ranging(LiDAR)PyTorch(Machine Learning Library)Confusion MatrixK-Means ClusteringUnsupervised LearningActivity

133、 RecognitionArtificial Intelligence DevelopmentMLflowPaddlePaddleGoogle AutoMLH2O.aiHugging Face(NLP Framework)Hugging Face TransformersDeepSpeechImage RecognitionOpenVINOMLOps(Machine Learning Operations)BERT(NLP Model)fastTextOpenAI GymKubeflowAIOps(Artificial Intelligence For IT Operations)Text-T

134、o-SpeechGPT-3(NLP Model)Sphinx Speech RecognitionExplainable AI(XAI)Generative Adversarial NetworksAI/ML InferenceMachine Learning Model Monitoring And EvaluationMachine Learning Model TrainingTransformer(Machine Learning Model)Variational AutoencodersChatGPTDeck.glLarge Language ModelingPyTorch Lig

135、htningAttention MechanismsBoltzmann MachineGenerative Artificial IntelligenceNuance MixPrompt EngineeringModelOpsOperationalizing AI30PwCs 2024 AI Jobs BarometerAppendix two:Calculating net skill changeHere is how the turnover in skills required by employers(net skill change)is calculated.How net sk

136、ill change is calculatedThe net skill change is a measure of the change in the frequency of skills required by employers for a particular occupation.This metric and its associated methodology to be calculated was developed by Harvard economists,David Deming and Kadeem Noray(2020).Below we present th

137、e formula and walk through an example.In short,the net skill change takes the absolute value of each skill change for an occupation and sums them.As it measures the absolute value the value is always positive.It is capturing skill changes be they positive or negative and adding them.The more changes

138、 in skills demanded by an employer be they demanded more or less(positive or negative),the higher this net skill change value.Formula:Example:If skill A is mentioned 50 times in 2019 and then 65 times in 2023(and we assume job postings remained constant in both time periods at 100 for example).The s

139、kill change would be 65/100-50/100=15/100=+0.15.If skill B is mentioned 30 times in 2019 and then 25 in 2023(in 100 postings in both periods),the skill change would be 25/100-30/100=-5/100=-0.05.The net skill change the sum of the absolute values:Net skill change for job X=0.15+0.05=0.20.2024 PwC.Al

140、l rights reserved.PwC refers to the PwC network and/or one or more of its member firms,each of which is a separate legal entity.Please see for further details.This content is for general information purposes only,and should not be used as a substitute for consultation with professional advisors.At P

141、wC,our purpose is to build trust in society and solve important problems.Were a network of firms in 151 countries with over 360,000 people who are committed to delivering quality in assurance,advisory and tax services.Find out more and tell us what matters to you by visiting us at .PwCs 2024 AI Jobs B