德勤：2024年Q2企業生成式人工智能報告（英文版）（31頁）.pdf

《德勤：2024年Q2企業生成式人工智能報告（英文版）（31頁）.pdf》由會員分享，可在線閱讀，更多相關《德勤：2024年Q2企業生成式人工智能報告（英文版）（31頁）.pdf（31頁珍藏版）》請在三個皮匠報告上搜索。

1、Deloittes State of Generative AI in the Enterprise Quarter two reportApril 2024Now decides next:Getting real about Generative AI findings1 Value creation2 Scaling up3 Building trust4 Evolving the workforceNext:Looking aheadAuthorship&AcknowledgmentsAbout the Deloitte AI Institute About the Deloitte

2、Center for Integrated Research About the Deloitte Center for Technology,Media&TelecommunicationsMethodologyTable of contents2 2IntroductionForewordWe have traveled a long way since the Generative AI space race kicked off in November 2022and yet,we know we are still at the beginning of this long and

3、exciting transformation.Every day,we talk with clients about how much there is to focus on in the moment,how explosive the pace of change is,and how challenging it can be amid the excitement to take a longer-term view.We see organizations starting to achieve benefits and move toward a near future wh

4、ere this early stage of Generative AI tools is widely dispersed and driving new value.But there are also some hard realities to deal with as business leaders look to scale and realize the potential of this powerful technology.The State of Generative AI in the Enterprise:Getting real about Generative

5、 AI captures a new snapshot of this transformative time from the perspectives of nearly 2,000 business and technology leaders,all from organizations that are actively deploying and scaling Generative AI today.Echoing our many clients,from these executives we hear that while excitement persists it ma

6、y be at its peak as leaders come up against cultural challenges,questions about how to manage their workforces,and issues with trust thatat least for nowstand in the way of unlocking Generative AIs full value.All told,it is exciting that Generative AIs potential is beginning to weave its way deeper

7、into the foundations of how organizations operate and we continue to learn more about emerging leading practices.Amid those developments,we also continue to see that achieving value with Generative AI connects hand in hand with keeping humans at the center.Learn more about the series and sign up for

8、 updates at http:/ Mittal,Costi Perricos,Kate Schmidt,Brenna Sniderman and David Jarvis“We are in the first inning of a thousand-inning game and theres so much to be figured out.”-Chief analytics officer in financial services3IntroductionGetting real about Generative AIIs the infatuation phase over?

9、Quarter two of Deloittes global quarterly survey found many organizations beginning to get down to the serious work of making Generative AIs vast potential a reality.This report presents findings from the second in Deloittes ongoing series of quarterly global surveys on Generative AI in the enterpri

10、se.To gain additional context for our wave two research,we also conducted a series of in-depth interviews with senior executives from a broad range of industries.Our research shows that organizations are increasingly prioritizing value creation and demanding tangible results from their Generative AI

11、 initiatives.This requires them to scale up their Generative AI deploymentsadvancing beyond experimentation,pilots and proofs of concept.Transitioning to large-scale deployments will increase Generative AIs impact on the business and expand its reach to a much larger segment of the workforce.Success

12、ful scaling,in turn,presents a wide range of challenges,encompassing everything from strategy,processes and people to data and technology.Two of the most critical challenges for scaling are building trust(in terms of making Generative AI both more trusted and trustworthy)and evolving the workforce(a

13、ddressing Generative AIs potentially massive impact on worker skills,roles and head count).Here well take an in-depth look at all four of these areasvalue,scaling,trust and workforceto help organizations move forward more effectively on their Generative AI journeys.Future survey reports will focus s

14、electively on other key challenges to successful Generative AI scaling and value creation.4Value creation The percentage of organizations reporting they were already achieving their expected benefits to a“large”or“very large”extent is 18%36%,depending on the type of benefit being pursued.Organizatio

15、ns that reported“high”or“very high”levels of Generative AI expertise are scaling Generative AI much more aggressivelyand are achieving their desired benefits to a much greater degree than others.Organizations primarily plan to reinvest the savings from Generative AI into innovation(45%)and improving

16、 operations(43%)addressing the value equation from both sides.Scaling up Leaders see scaling as essential for creating value,increasing Generative AIs impact on the business and expanding the technologys user base.The scaling phase is when Generative AIs potential benefits are converted into real-wo

17、rld value.Its also,however,when an organizations potential concerns can become real-world barriers to success.Common areas of concern include data security and quality,explainability of Generative AI outputs,and worker mistrust or lack of familiarity with Generative AI tools.Workforce access to appr

18、oved Generative AI tools and applications remains quite low,with nearly half of surveyed organizations(46%)reporting they provided approved Generative AI access to just a small portion of their workforces(20%or less).However,most workers with internet access will have access to public Generative AI

19、tools and could be using them without consent.IntroductionGetting real about Generative AI(contd)All statistics noted in this report and its graphics are derived from Deloittes second quarterly survey,conducted January February 2024;The State of Generative AI in the Enterprise:Now decides next,a rep

20、ort series.N(Total leader survey responses)=1,982.Generative AI is an area of artificial intelligence and refers to AI that in response to a query can create text,images,video and other assets.Generative AI systems can interact with humans and are often built using large language models(LLMs).Also r

21、eferred to as“GenAI.”5Building trust Lack of trust remains a major barrier to large-scale Generative AI adoption and deployment.Two key aspects of trust we observed are:(1)trust in the quality and reliability of Generative AIs output and(2)trust from workers that the technology will make their jobs

22、easier without replacing them.Trust issues have not prevented organizations from rapidly adopting Generative AI for experiments and proofs of concept,with 60%reporting they are effectively balancing rapid implementation with risk management.Trust is likely to become a bigger issue,however,as organiz

23、ations transition to large-scale deployment.Many reported they are currently investing significant time and effort into building guardrails around Generative AI.Organizations that reported“high”or“very high”levels of expertise recognize the importance of building trust in Generative AI across numero

24、us dimensions(e.g.,input/output quality,transparency,worker empathy)and are implementing processes to improve it to a much greater extent than are other organizations.Evolving the workforce Most organizations(75%)expect the technology to affect their talent strategies within two years;32%of organiza

25、tions that reported“very high”levels of Generative AI expertise are already making changes.The most expected talent strategy impacts are process redesign(48%)and upskilling or reskilling(47%).Generative AI is expected to increase the value of some technology-centered skills(such as data analysis)as

26、well as human-centered skills(such as critical thinking,creativity and flexibility),while decreasing the value of other skills.In the short term,more organizations said they expect the technology to increase head count(39%)than to decrease head count(22%)perhaps due to increased needs for Generative

27、 AI and data expertise.The wave two survey covered in this report was fielded to 1,982 director-to C-suite-level respondents across six industries and six countries between January and February 2024.Industries included:Consumer;Energy,Resources&Industrials;Financial Services;Life Sciences&Health Car

28、e;Technology,Media&Telecom;and Government&Public Services.Our Q2 survey findings are augmented with over 20 executive interviews.This second report is part of a yearlong series by the Deloitte AI Institute to help leaders in business,technology and the public sector track the rapid pace of Generativ

29、e AI change and adoption.The series is based on Deloittes State of AI in the Enterprise reports,which have been released annually the past five years.Learn more at the State of Generative AI in the Enterprise:Wave two survey resultsIntroductionGetting real about Generative AI(contd)6 Now:Key finding

30、s7 7Proving,measuring and communicating value is crucial to an organizations Generative AI journey.In our survey and interviews,many organizations reported they were increasingly emphasizing the need for Generative AI initiatives and investments to have clear value objectives and deliver tangible re

31、sults,rather than simply being viewed as experiments or learning experiences.As one executive at a Fortune 500 manufacturing company noted:“We have a very strict internal rule that if we dont see value from our Generative AI solutions,we wont do it or we wont scale it.”That said,there are many ways

32、to define and measure valueespecially for a technology with the transformational potential of Generative AI.Although financial return on investment(ROI)is important,value drivers such as innovation,strategic positioning and competitive differentiation can be even more important.Value objectives and

33、priorities for Generative AI canand shouldvary by organization,industry and use case.Where the technologys potential impact is strategic and truly game-changing,the need and latitude for experimentation,learning and innovation are much greater(with less emphasis on immediate payback)than in situatio

34、ns where productivity and cost savings are the primary expected benefits.Moreover,Generative AI is so newand advancing so quicklythat accurately estimating benefits is much harder than for an established technology with a proven track record.“Any technology thats a little over a year old,nobodys goi

35、ng to have a years worth of data to do a backward-looking ROI,”said one tech company executive we interviewed.“And with the fundamental and foundational changes Generative AI offers,its very hard to even offer a forward-looking total cost of operating or ROI because theres so many possibilities of i

36、mpact and varied ways to integrate it into your business.”Therefore,many forward-thinking organizations are implementing Generative AI without specific ROI targets as they realize they cant afford to get left behind in this critical and fast-moving market.1Value creationNow:Key findings8 8Organizati

37、ons are starting to demand tangible business value from Generative AI,and some are beginning to achieve real-world benefits.The organizations we surveyed expect Generative AI to deliver a broad range of benefits,with the most common objectiveat least in the short termbeing improved efficiency and pr

38、oductivity(56%),which is consistent with the results from last quarters survey.The percentage of respondents who said their organizations Generative AI initiatives were already achieving expected benefits to a“large”or“very large”extent is 18%36%,depending on the type of benefit being pursued.As one

39、 public sector executive told us,“The big selling point is if I make an investment and do something like this,whats the tangible return and what are some easy returns?And then what are more complicated longer-term returns that take more investment money?If I can do some of the easier ones and build

40、on them,it can translate into I think this would be worth it to invest a lot more money.I believe a lot of entities in our sector are at that point.”Generative AI“experts”are achieving their desired benefits to a much greater degree.In every category,organizations that rated themselves as having“hig

41、h”or“very high”levels of Generative AI expertise reported much greater success at achieving their desired benefits.Their advantage was greatest in strategic and growth-related areas such as improving products and services and encouraging innovation and growth.Now:Key findingsQ:What are your anticipa

42、ted benefits and to what extent are you achieving those benefits to date?(Jan./Feb.2024);N(Total)=1,982;N(very high)=96Figure 1Encourage innovation and growthImprove efficiency and productivityReduce costsIncrease speed/ease of dev new systems/softwareUncover new ideas and insightsIncrease revenueEn

43、hance relationships with clients/customersImprove existing products and servicesDetect fraud/manage riskShift workers from lower-to higher-level tasksVery high expertiseAchieving benefitsOverall28%35%27%36%30%25%29%30%22%70%63%54%55%48%40%36%48%48%42%18%Of those seeking the benefit,the percentage of

44、 respondents achieving the benefit to a large extent or more9“Expert”organizations are scaling Generative AI much more aggressively.Generative AI expert organizations are likely having more success at capturing benefits because they are scaling up much more aggressively,compared to the other categor

45、ies,which provides a larger base for generating benefits.According to our survey,organizations reporting“very high”levels of Generative AI expertise are deploying AI much more rapidly and extensively than others.In fact,73%said they are adopting the technology at a“fast”or“very fast”pace(versus only

46、 40%of organizations with“some”level of expertise).They are also scaling Generative AI at higher rates across functions and using it more within functions.For example,those with“very high”expertise reported,on average,implementing at scale in 1.4 functions,out of eight total functions,while those wi

47、th“some”expertise are doing so in only 0.3 functions.Further,38%of those with“very high”expertise reported implementing Generative AI at scale in marketing,sales and customer serviceversus only 10%of organizations with“some”level of expertise.Figure 2Adopting at a faster pace Adopting Generative AI“

48、fast”or“very fast”73%40%Providing more of their workforce access to GenAI 40%of workforce has access to Generative AI tools/applicationsAdopting at higher levels across functions Implementing Generative AI for marketing,sales and customer serviceInvesting more in hardware Increasing hardware investm

49、ent because of Generative AI strategyInvesting more in cloud consumption Increasing cloud investment because of Generative AI strategyUsing code generators more Currently using Generative AI code generatorUsing open-source LLMs more Currently using open source large language models48%23%66%33%61%39%

50、80%62%64%34%47%19%(Jan./Feb.2024)N(Total)=1,982;N(Very high)=96;N(Some)=1,021Very high expertiseSome expertiseNow:Key findingsCompanies that report expertise are moving quickly.10Insights from our executive interviews align closely with survey findings,showing that leading organizations are aggressi

51、vely scaling up their Generative AI efforts both horizontally(across multiple functions or domains)and vertically(within a single function or domain).This combination of horizontal and vertical scaling may help achieve value creation more effectively.As one chief transformation officer in manufactur

52、ing noted,“We have an application that is being incredibly successful and has saved us significant amounts of money and that we have scaled very broadly across many of our sites and continue to scale further across more equipment across more sites.”Similarly,from a broad market perspective we are se

53、eing an increasingly sharp distinction between horizontal use cases that cut across industries(e.g.,office productivity suites and enterprise resource planning systems with integrated Generative AI)and vertical use cases that are industry-specific and narrowly focused but more strategically impactfu

54、l(e.g.,Generative AI tools for semiconductor design that are used only by a small subset of workers but have a very large impact on the business).Now:Key findings11Now:Key findingsOrganizations primarily plan to reinvest the savings from Generative AI into innovation and additional operations improv

55、ements.Among the overall respondent pool,organizations said they primarily planned to reinvest cost and timesavings from Generative AI into driving innovation(45%)and improving operations(43%),addressing the value equation from both sides.Its interesting to note that a significant percentage of orga

56、nizations(27%)also planned to reinvest in scaling Generative AI adoption,creating a cycle of Generative AI reinvestment and growth.Organizations with“very high”Generative AI expertise are even more focused than others on driving innovation(51%).They are also less inclined than others to reinvest sav

57、ings from Generative AI into improving operations and more inclined to prioritize developing new products and services.The right reinvestment approach depends on an organizations specific needs.Organizations currently facing strategic disruption or transformation from Generative AI have a greater im

58、perative to focus on strategic objectives such as innovation and growth,and are likely already working more aggressively to develop strong Generative AI capabilities.By contrast,organizations in industries that are currently not being disrupted by Generative AI are more likely to focus on benefits s

59、uch as individual worker productivity and operations improvement,areas with less of a sense of urgency and less tolerance for risk.Such organizations can still benefit greatly from Generative AIjust in a different way.They also have a valuable opportunity to watch and learn from the experiences of o

60、ther industries that are currently being disruptedlessons that could serve them well if and when Generative AI disruption reaches their own industry.“To enable GenAI value in our business,we need to change our mindset and develop R&D capabilities to realize a long-term vision enabled by GenAI,”said

61、the CEO of a digital media company.“Right now,our mindset is short-term and just about tangible cash value for one-off use cases.”9%16%13%19%Exploring new business models45%Driving innovation opportunities43%Improving operations across the organizationTraining and upskilling employeesImproving cyber

62、security infrastructureEnhancing IT infrastructureEnhancing risk management systems29%Developing new products and servicesExpanding our market27%Scaling GenAI adoption across the organizationCreating new jobsCreating a return for shareholders28%23%20%28%Areas to reinvest time and cost savingsFigure

63、3Q:Where does your company plan to reinvest cost or timesavings generated through implementation of GenAI capabilities(select top 3)?(Jan./Feb.2024)N(Total)=1,98212Scaling upNow:Key findingsA key to value creation,scaling increases Generative AIs impact on the business and expands its user baseboth

64、of which have a strong multiplier effect on Generative AIs benefits.Yet,many organizations find it challenging to make the leap from pilots and proofs of concept to large-scale deployment.Scaling is complex and requires effort across a variety of interrelated elements spanning strategy,process,peopl

65、e,data and technology.Although the challenges associated with scaling Generative AI are common to many digital transformation initiatives,issues such as risk management and governance,workforce transformation,trust and data management take on even greater importance.What worked well in the past migh

66、t not work the same way with this new technology.The scaling phase is when potential benefits are converted into real-world value.It is also,however,when potential issues become real-world barriers.And with Generative AI,many of those barriers are still being identified and understood.“There are alw

67、ays issues that emerge through the adoption and scaling transition that arent expectedthe question we have to consider is how hard are they to overcome,”said a chief technology officer we interviewed.“For example,one of our use cases had some technical,policy and cybersecurity issues,but they were r

68、elatively easy to overcome,so we scaled.Conversely,for two other use cases more issues emerged linked to the skill level to work with the outputs of the AI solution.These have been harder to address,so scaling has been slower.”A public sector chief information officer outlined another approach:“For

69、us,successful scaling is building on previous successes and then taking those initiatives to another level.Expanding to other areas of the organization,incorporating more datasets,expanding the user base(internal and external)to improve upon existing results,and refining the current solution for mor

70、e value.This phased approach gives us a sense of assurance the investment is worthwhile before we commit significantly more resources.”Off-the-shelf Generative AI solutions for common use cases such as office productivity are arguably the easiest to deploy at scale,but they still require substantial

71、 investment,effort and training.For unique and/or more strategic Generative AI solutions and use cases,the complexity and challenges increase by leaps and bounds,along with the potential for greater returns.213Workforce access to approved GenAI tools and applications remains low.Nearly half of our r

72、espondents(46%)reported they provided approved Generative AI access to just a small portion of their workforces(20%or less).Organizations reporting “very high”levels of Generative AI expertise are further along,with nearly half(48%)providing approved Generative AI access to at least 40%of their work

73、forces.Even for these“expert”organizations,worker access to approved tools remains the exception,not the rule.Our executive interviews pointed to a number of reasons for this overall low penetration rate,mostly revolving around risk versus rewardespecially data-related risks.Do the potential rewards

74、 of Generative AI justify the risks,and can the risks be mitigated?In particular,we found widespread concern that allowing workers to use public large language models(LLMs)and Generative AI tools might lead to problems with protection of intellectual property and customer privacy.Now:Key findingsFig

75、ure 4Percentage of workforce with access to Generative AIQ:How much of your overall workforce,do you estimate,have access to your organizations sanctioned(approved)Generative AI tools/applications?(Jan./Feb.2024)N(Total)=1,982,N(Very high)=96,N(High)=606,N(Some)=1,021,N(Little)=257OverallLittle expe

76、rtiseSome expertiseHigh expertiseVery high expertiseUp to 20%20%40%40%60%60%80%More than 80%Percentage of the workforce46%76%49%31%25%29%16%28%36%27%16%5%14%23%24%3%1%3%4%8%6%2%6%7%16%14Now:Key findingsOther concerns that came up in our executive interviews include:Generative AI outputs that can be

77、unpredictable and subject to inaccuracies(i.e.,“hallucinations”)which undermine trust,particularly when combined with lack of transparency and explainability Potential loss of control over what Generative AI apps are being used within the organization and who is using them Worker resistance to using

78、 Generative AI due to lack of familiarity or concerns about being replacedGiven the potential challenges and risks,a cautious approach to allowing workers to use Generative AI tools arguably makes sense.However,tight restrictions on Generative AI are best viewed as a temporary stopgap measurenot a v

79、iable long-term solution.Logically,any worker with internet access will have access to public Generative AI tools and could be using them without their employers consentpotentially leaking sensitive data and intellectual property into public LLMs in an entirely uncontrolled way.This status is likely

80、 to continue in the absence of practical policies for allowing and managing widespread Generative AI access.Organizations should be actively developing sustainable processes and policies for enabling ubiquitous but responsible Generative AI use and managing the associated risks at scale.Widespread b

81、ut controlled access to Generative AI will help people get more comfortable with the technology and enable them to understand what it can and cannot dogiving them a more realistic and informed perspective while opening the door to new opportunities for Generative AI value creation across the enterpr

82、ise.1515“It has been surprising to see how low the bar is to do something quick and dirtythis is both exciting and scary,but the big challenge is to scalethis is a whole new ball game but scaling is hard without centralization.”-Director of data science and AI in the technology industry1616Building

83、trustNow:Key findings3Growing trustsaid their organizations trust in all forms of AI has increased since Generative AI emerged in late 202272%Measuring trustof organizations said they are measuring worker trust and engagement as part of altering their talent strategies because of the adoption of Gen

84、erative AI36%Lack of trust continues to be one of the biggest barriers to large-scale adoption and deployment of Generative AI.In this context,two key aspects of trust are:(1)trust in the quality and reliability of Generative AIs output(supported by improved transparency and explainability),and(2)tr

85、ust from workers that Generative AI will make their jobs easier and wont replace them.Regarding worker trust,one executive we interviewed noted that“once people start seeing efficiencies and the benefits the tools have to their work,that will drive adoption and sustained success.”In other words,grea

86、ter exposure to Generative AI tools will help people become more comfortable with the technology and understand how it can help them do their jobs.As for trusting Generative AIs outputs,the technologys fallibility in the form of“hallucinations”is well known and is actively being addressed through im

87、proved training and guardrails.For many organizations,transparency and explainability are even bigger issues.In its current form,Generative AI still operates largely as a black boxtaking an input and producing an output with no real way for humans to understand how that output was reached.According

88、to a chief technology officer we interviewed,“The explainability piece is really holding us back right now once we get a better handle on that,I think we will really be able to accelerate our adoption.”Ultimately,most organizations will likely each end up using LLMs customized and fine-tuned for the

89、ir specific domain,industry and use case,rather than just scaling up a general-purpose LLM.This specificity will help Generative AI produce outputs that are more precise,transparent and explainable.Lack of trust and related risks have thus far not prevented organizations from rapidly adopting Genera

90、tive AI for experiments and proofs of concept;however,this will likely change as organizations transition to large-scale deployment.According to our wave two survey,60%of respondents believed their organization is effectively balancing rapid integration of Generative AI while implementing processes

91、that mitigate potential risks.Also,72%said their organizations trust in Generative AI has increased since the technology emerged in late 2022.Lacking confidenceselected“lack of confidence in results”as one of their top risks related to Generative AI tools/applications(#3 of 10 overall)33%Balancing s

92、peed and riskreported their organization is effectively balancing integrating Generative AI rapidly while implementing processes that mitigate potential risks60%Opportunities for improvementAreas of strengthFigure 5(Jan./Feb.2024)N(Total)=1,98217Our executive interviews suggest,however,that addressi

93、ng trust issues is likely to become critically important as organizations transition from experimentation to large-scale deploymentespecially for organizations where the imperative to deploy Generative AI is more tactical than strategic,and thus less time sensitive.Generative AI when deployed at sca

94、le becomes far more important to the business and affects a much larger pool of human users,making trust a much bigger issue.Trust related to data quality,LLM training and output reliability becomes particularly important.“If you dont have the right dataset or data quality,it is very hard for the ap

95、plication to be helpful,”said a chief technology officer we interviewed.“GenAI solutions are very sensitive to good quality and well-structured data.If the data is not correct,it is very hard to know that the output is wrong.”In our survey,33%of respondents cited lack of confidence in results as one

96、 of Generative AIs top risks(third in the list of top 10 risks).Only 36%of the organizations surveyed were measuring worker trust and engagement as part of adapting their talent strategies to Generative AI.“Expert”organizations recognize the importance of building trust in Generative AI and are putt

97、ing effort into it.Despite the importance of trust for successful Generative AI deployment and scaling,40%45%of our overall respondents said they are,to a“large”or“very large”extent,implementing processes to improve trust in their Generative AI initiatives through various aspects(such as data qualit

98、y,output reliability and organizational empathy).However,among organizations that reported“very high”Generative AI expertise,the focus on trust is much higher across every aspect(59%73%).This likely reflects both their greater appreciation for the importance of trust and their greater reliance on Ge

99、nerative AI as an integral and crucial part of the business.Now:Key findingsOverallVery high expertiseTransparency with employees45%60%40%59%43%73%41%67%Demonstration of consideration,empathy and kindness in use of GenAIQuality Generative AI input dataReliable Generative AI outputFigure 6(Jan./Feb.2

100、024)N(Total)=1,982;N(Very high)=96Companies implementing processes to generate trust in GenAITo a“large”or“very large”extent18Evolving the workforceNow:Key findingsWorkforce challenges affect Generative AI scaling on both the front and back ends.On the front end,organizations need valuable and scarc

101、e talent with expertise in Generative AI(and data management)to develop and refine their solutions.They also need the overall workforce to be comfortable enough with the technology to be willing to use it for improving efficiency and effectiveness.On the back end,organizations need to understand how

102、 the workforce could be affected by large-scale Generative AI deployment and then develop talent strategies,programs and policies that make sense for the business and workers alike.Addressing these critical and complex workforce issues is an urgent enabler for Generative AI adoption and scaling,even

103、 as organizations work to figure out the technology side of the problem.Most organizations expect Generative AI to affect their talent strategies.Three-quarters of survey respondents(75%)expect to change their talent strategies within two years in response to Generative AI.Organizations reporting“ve

104、ry high”Generative AI expertise expect to change their talent strategies even faster,with 32%already making changes.This is consistent with our broader finding that such organizations are scaling up their initiatives much more aggressively than are others,leading to greater and more immediate talent

105、 impacts.418%26%31%16%10%NowWithin 1 year1-2 years2+yearsDont know/no formal time frameTimeline for change in talent strategiesQ:When do you expect to make changes in talent strategies because of generative AI?(Jan./Feb.2024)N(Total)=1,982Figure 719The most common talent strategy responses are proce

106、ss redesign and upskilling or reskilling.In response to Generative AI adoption,the most common changes to talent strategy among the overall respondent pool involve redesigning work processes(48%)and upskilling or reskilling(47%).Relative to the overall respondent pool,organizations with“very high”ex

107、pertise were much more focused on developing AI fluency(47%)and redesigning career paths(38%),and much less focused on assessing changes to the anticipated supply and demand of skills(25%).These survey results suggest a strong need for more attention paid to Generative AIs talent impacts.In the near

108、 term,AI education and fluency will be especially important to fostering adoption and overcoming initial resistance to change.In the longer term,upskilling or reskilling and redesigning work processes and career paths will likely be essential for capturing Generative AIs full value and positioning w

109、orkers for future success.As one executive noted,“In general,I think its more about upskilling the people you have,because whats really valuable is the domain knowledge and the relationships and all that.”Now:Key findingsHow companies are adjusting talent strategiesFigure 8Q:How is your organization

110、 adjusting its talent strategies because of the adoption of generative AI?(Jan./Feb.2024)N(Total)=1,98229%48%Redesigning work processes to take advantage of Generative AI47%Designing and implementing upskilling and reskilling strategiesRedesigning career paths and career mobility strategiesAssessing

111、 target talent acquisition levels and workforce ecosystem access strategyProviding performance-based incentives for leveraging Generative AI36%Measuring worker trust and engagementLaunching AI fluency development programs35%Assessing changes to the anticipated supply and demand of skills29%35%34%20G

112、enerative AI is expected to increase the value of certain technology-centered and human-centered skills,while decreasing the value of others.In the Generative AI era,the competencies that organizations require from their workforces will evolve.An emphasis on new technical skills will combine with a

113、renewed focus on the skills that make people uniquely human and valuable.According to the survey results:Technology-centered skills that respondents most expect to increase in value include:data analysis(70%),prompt engineering(60%),information research(59%),and software engineering/coding(57%).Huma

114、n-centered skills that respondents most expect to increase in value include:critical thinking and problem-solving(62%),creativity(59%),flexibility/resilience(58%),and the ability to work in teams(54%).The value of specific skills will likely vary depending on timing and organizational level.A key ch

115、allenge for todays organizations is figuring out how to help workers harness the power of Generative AI to do their jobs more efficiently and effectively and create more value for the business.Now:Key findingsSkills rising in valueMore or much more valuableFigure 9Human-centered skillsTechnology-cen

116、tered skillsManual content creation31%51%51%53%Communication skills54%Ability to work in teamsData analysisPrompt engineeringInformation research57%Software engineering/codingApplication developmentEmotional intelligenceCritical thinking/problem-solvingCreativityFlexibility/resilience59%59%60%62%70%

117、58%Q:To what extent are the following workforce skills going to be more or less valuable across your organization because of the adoption of generative AI tools/capabilities?(Jan./Feb.2024)N(Total)=1,98221Head count is expected to increase slightly (at least in the short term).The survey results sho

118、w that more organizations expect to increase head count(39%)than to decrease head count(22%)over the next 12 months due to implementation of their Generative AI strategy.This is especially true for organizations with“very high”Generative AI expertise(45%increase in head count vs.23%decrease in head

119、count)or high expertise(46%increase vs.25%decrease).Our executive interviews present a cloudier picture on long-term head count.A widespread focus on value creation through efficiency and productivity improvement implies that organizations are striving to do more with less,which could lead to head c

120、ount reduction or reduced hiring.Now:Key findingsLittle expertise14%57%27%Some expertise21%39%38%High expertise25%29%46%Very high expertise23%28%45%Overall enterprise head count changeDont knowDecreaseNo changeIncreaseFigure 10Q:Which of the following best describes the full-time head count change y

121、ou anticipate will result over the next 12 months due to implementation of your organizations generative AI strategy?(Jan./Feb.2024)N(Total)=1,982;N(Very high)=96;(High)=606;N(Some)=1,021;N(Little)=257Head count change by enterprise expertise levelIncrease significantly1%Increase moderately6%Increas

122、e slightly32%No change38%Decrease moderately3%Decrease significantly(by more than 20%)1%Dont know2%Decrease slightly(up to 10%)18%(by 10%20%)(by more than 20%)(up to 10%)(by 10%20%)22On the other hand,many breakthrough technologies in human history have raised the specter of widespread job displacem

123、entyet society has always found new and valuable ways to employ its human capital.Will the impact of AI be fundamentally different?Maybe.Maybe not.One thing that seems certain,however,is that some roles and skills will be more affected than others.Although Generative AIs net impact on employment at

124、the societal level might be neutral or even positive,the impact for affected individuals at the personal level could be profoundly challenging.“GenAI is already having a significant impact on our talent requirement perspectiveit has already had an impact on our head count and we are now looking for

125、different talent than we did in the past,”said a Fortune 500 executive in the manufacturing industry we interviewed.“The tipping point is coming and we are not far away from this:Once we can truly scale process improvement and plug GenAI into the product life cycle,this would be a major shift.”Most

126、organizations acknowledge they are still in the early stages of adapting their talent and HR strategies to Generative AIs impactsand they recognize the need to focus on them sooner rather than later.“Best-case GenAI scenario for us is that we have high-quality jobs that create a differentiating plac

127、e to work relative to our competitors that helps us attract and retain talent,”said a chief analytics officer we interviewed.“The worst case is we are so behind that we cant hire,our products are inferior,and we took a risk on something that created a significant legal issue that is impacting our bu

128、siness.”Now:Key findings2323 Next:Looking ahead2424In the short term,most organizations primarily view Generative AI as a tool to improve productivity at the individual level and efficiency at the functional level.And while some organizations are starting to see tangible results in those areas,our s

129、urvey findings show that organizations reporting higher levels of Generative AI expertise tend to be more focused than others on innovation and developing new products and services.Dont let the uncertainty of this moment stop you from imagining a fundamentally different future for your organization.

130、In the long term,we believe the big winners will use the technology to differentiate themselves,enable broad enterprise transformation and create value in new ways.Efficiency is goodinnovation and growth are better.Next:Looking ahead25To succeed at scaling,you will likely need to concurrently evolve

131、 your strategy,processes,people,data and technology.Also,your organization will probably need to develop strong capabilities for both horizontal and vertical scalingmeaning bringing Generative AI capabilities to as many workers as possible while at the same time deeply embedding other capabilities i

132、n specific functions or processes.Establishing a center of excellence for Generative AI can help.It should provide centralized resources(teams,tools,processes,policies,knowledge and experience)that can accelerate deployment of similar use cases and enable you to make the most of scarce expertise.Mor

133、e broadly,organizations need to invest in the foundations of Generative AI:data modernization,talent,and technology and infrastructure.These foundational investments will likely deliver value across multiple projectsand across the entire enterprise.Most important:dont wait for rock-solid proof of fi

134、nancial benefits before starting on scaling.Although it might require an organizational leap of faith,the way to maximize the value of Generative AI is to advance from proofs of concept to full implementations.Tackle barriers to scaling.As Generative AI moves from possibilities to practicalities,cho

135、osing the right use cases,selecting appropriate tools,getting to scale,and accurately measuring progress will all be important steps.A holistic approach to value realizationboth financial and nonfinancialis vital.Having the right processes in place to measure all value created by Generative AI will

136、help you determine if you are achieving value in unexpected ways.Communicating the value created by Generative AI will also be critical to helping build momentum and support for continued progress.Proving,measuring and communicating ALL types of value is critical.Next:Looking ahead26In todays extrem

137、ely competitive market for AI talent,organizations are not only aggressively pursuing new talent,but also training their broader workforces on Generative AI.When Generative AI is deployed at scale,different technical and human-focused skills will become more important.To succeed,organizations will l

138、ikely need to move beyond simple fluency and create new roles,new work processes,and a new organizational culturewith an active focus on developing junior talent into senior talent that can use Generative AI to its full advantage.Organizations might also need to realign their existing internal resou

139、rces around projects,which could include centralizing Generative AI talent to simultaneously support multiple initiatives across the enterprise.To scale up,you need to skill up.Trust is the foundation for increased adoption.Without it,widespread use of Generative AI wont happen.Improving AI fluency

140、and providing broader access to Generative AI tools can help people get more comfortable with the technology and gain a more realistic perspective on what it can and cannot do.Having the right data management,technology infrastructure and governance in place to help ensure high-quality inputs as wel

141、l as verified and explainable outputs will also help build trust.As a leader,you can also actively instill trust in Generative AI throughout your organization by(1)clearly and regularly communicating your strategic objectives for Generative AI,(2)fostering a culture of curiosity that encourages empl

142、oyees to experiment with the new tools,and(3)frequently measuring worker trust to uncover potential frustrations and barriers to adoption.Build trust through transparency,familiarity,technology and guardrails.Next:Looking ahead27Brenna Sniderman Executive Director Deloitte Center for Integrated Rese

143、arch Deloitte Services LLP Authorship and AcknowledgmentsNitin Mittal Global AI Leader Deloitte LLP David Jarvis Senior Research Leader Deloitte Center for Technology,Media&Telecommunications Deloitte Services LP Acknowledgments We would like to thank our leaders and many talented professionals who

144、brought this research to life:Beena Ammanath,Deborshi Dutt,Kevin Westcott,Lynne Sterrett and Jeff Loucks;Ahmed Alibage,Eric Alons-Cruz,Siri Anderson,Sean Benton,Natasha Buckley,Amber Bushnell,Maria Fernanda Castro,Tracy Fulham,Jordan Garrick,Gerson Lehrman Group(GLG),Lou Ghaddar,Jessi Hendon,Tatum H

145、oehn,Karen Hogger,Jonathan Holdowsky,Lisa Iliff,Justin Joyner,Lena La,David Levin,Michael Lim,Nina Lukina,Joe Mariani,Rajesh Medisetti,Sharonjeet Meht,Judy Freeman Mills,Melissa Neumann,Jamie Palmeroni-Lavis,Jose Porras,Jonathan Pryce,Negina Rood,Lesley Stephen,Kelcey Strong,10 EQS,Sandeep Vellanki,

146、Ivana Vucenovic,Marianne Wilkinson and Sourabh Yaduvanshi.We would also like to thank additional Deloitte subject matter specialists who contributed to the development of the survey and report:Ed Bowen,Bjoern Bringmann,Lou DiLorenzo,Amelia Dunlop,Maggie Fletcher,Rohan Gupta,Oz Karan,Kellie Nuttal,El

147、izabeth Powers,Ashley Reichheld,Dany Rifkin,Jim Rowan,Baris Sarer,Mike Segala,Laura Shact,Rohit Tandon,Ed Van Buren and Greg Vert.Costi Perricos Global Office of Generative AI Leader Deloitte UK cperricosdeloitte.co.ukKate Schmidt Chief Operating Officer Deloitte AI Strategic Growth Offering Deloitt

148、e Consulting LLP Business leadershipResearch leadership28About the Deloitte AI Institute The Deloitte AI Institute helps organizations connect all the different dimensions of the robust,highly dynamic and rapidly evolving AI ecosystem.The AI Institute leads conversations on applied AI innovation acr

149、oss industries,using cutting-edge insights to promote human-machine collaboration in the Age of With.The Deloitte AI Institute aims to promote dialogue about and development of artificial intelligence,stimulate innovation,and examine challenges to AI implementation and ways to address them.The AI In

150、stitute collaborates with an ecosystem composed of academic research groups,startups,entrepreneurs,innovators,mature AI product leaders and AI visionaries to explore key areas of artificial intelligence including risks,policies,ethics,future of work and talent,and applied AI use cases.Combined with

151、Deloittes deep knowledge and experience in artificial intelligence applications,the institute helps make sense of this complex ecosystem and,as a result,delivers impactful perspectives to help organizations succeed by making informed AI decisions.About the Deloitte Center for Integrated ResearchThe

152、Deloitte Center for Integrated Research(CIR)offers rigorously researched and data-driven perspectives on critical issues affecting businesses today.We sit at the center of Deloittes industry and functional expertise,combining the leading insights from across our firm to help leaders confidently comp

153、ete in todays ever-changing marketplace.About the Deloitte Center for Technology,Media&TelecommunicationsThe Deloitte Center for Technology,Media&Telecommunications(TMT Center)is a world-class research organization that serves Deloittes TMT practice and our clients.Our team of professional researche

154、rs produce practical foresight,fresh insights,and trustworthy data to help clients see clearly,act decisively and compete with confidence.We create original research using a combination of rigorous methodologies and deep TMT industry knowledge.Learn moreLearn moreLearn more29To obtain a global view

155、of how Generative AI is being adopted by organizations on the leading edge of AI,Deloitte surveyed 1,982 leaders between January and February 2024.Respondents were senior leaders in their organization and included board and C-suite members,and those at the president,vice president and director level

156、.The survey sample was split equally between IT and line of business leaders.Six countries were represented:Australia(99 respondents),France(131 respondents),Germany(150 respondents),India(200 respondents),the United Kingdom(200 respondents),and the United States(1,202 respondents).All participating

157、 organizations have one or more working implementations of AI being used daily.Plus,they have pilots in place to explore Generative AI or have one or more working implementations of Generative AI being used daily.Respondents were required to meet one of the following criteria with respect to their o

158、rganizations AI and data science strategy,investments,implementation approach and value measurement.They:influence decision-making,are part of a team that makes decisions,are the final decision-maker,or manage or oversee AI technology implementations.All statistics noted in this report and its graph

159、ics are derived from Deloittes second quarterly survey,conducted January February 2024;The State of Generative AI in the Enterprise:Now decides next,a report series.N(Total leader survey responses)=1,982Methodology30About DeloitteDeloitte refers to one or more of Deloitte Touche Tohmatsu Limited(DTT

160、L),its global network of member firms,and their related entities(collectively,the“Deloitte organization”).DTTL(also referred to as“Deloitte Global”)and each of its member firms and related entities are legally separate and independent entities,which cannot obligate or bind each other in respect of t

161、hird parties.DTTL and each DTTL member firm and related entity is liable only for its own acts and omissions,and not those of each other.DTTL does not provide services to clients.Please see to learn more.Deloitte provides industry-leading audit and assurance,tax and related services,consulting,finan

162、cial advisory,and risk advisory services to nearly 90%of the Fortune Global 500 and thousands of private companies.Our people deliver measurable and lasting results that help reinforce public trust in capital markets,enable clients to transform and thrive,and lead the way toward a stronger economy,a

163、 more equitable society,and a sustainable world.Building on its 175-plus year history,Deloitte spans more than 150 countries and territories.Learn how Deloittes approximately 457,000 people worldwide make an impact that matters at .This publication contains general information only and Deloitte is n

164、ot,by means of this publication,rendering accounting,business,financial,investment,legal,tax,or other professional advice or services.This publication is not a substitute for such professional advice or services,nor should it be used as a basis for any decision or action that may affect your business.Before making any decision or taking any action that may affect your business,you should consult a qualified professional advisor.Deloitte shall not be responsible for any loss sustained by any person who relies on this publication.Copyright 2024 Deloitte Development LLC.All rights