《埃森哲&WEF:2024未来工作白皮书:大语言模型与就业:商业工具包(英文版)(17页).pdf》由会员分享,可在线阅读,更多相关《埃森哲&WEF:2024未来工作白皮书:大语言模型与就业:商业工具包(英文版)(17页).pdf(17页珍藏版)》请在三个皮匠报告上搜索。
1、Jobs of Tomorrow Large Language Models and Jobs A Business ToolkitW H I T E P A P E RD E C E M B E R 2 0 2 3In collaboration with AccentureImages:Getty Images 2023 World Economic Forum.All rights reserved.No part of this publication may be reproduced or transmitted in any form or by any means,includ
2、ing photocopying and recording,or by any information storage and retrieval system.Disclaimer This document is published by the World Economic Forum as a contribution to a project,insight area or interaction.The findings,interpretations and conclusions expressed herein are a result of a collaborative
3、 process facilitated and endorsed by the World Economic Forum but whose results do not necessarily represent the views of the World Economic Forum,nor the entirety of its Members,Partners or other stakeholders.ContentsForeword 3Executive summary 4Introduction 51 Consideration 1:Job change and job di
4、splacement risk 71.1 Responsible business strategies 82 Consideration 2:Job quality 92.1 Responsible business strategies 103 Consideration 3:Skilling and learning 113.1 Responsible business strategies 12Conclusion 13Contributors 14Endnotes 15Jobs of Tomorrow Large Language Models and Jobs A Business
5、 Toolkit2ForewordGenerative artificial intelligence(AI)and,in particular,large language models(LLMs),underpinned by advancements in machine learning and natural language processing,represent a paradigm shift in how we interact with information and,by extension,how we work.These technologies can crea
6、te original content,generate insights from large amounts of data,translate languages with near-human accuracy and potentially even make complex decisions.The versatility and efficiency of these technologies,including new LLM-powered human-machine interfaces such as intelligent agents,could have prof
7、ound implications for jobs and the future of work.While the application of LLMs could lead to significant productivity gains and the creation of new types of jobs,there is also a risk that they could displace existing roles,exacerbating socioeconomic disparities and creating a sense of job insecurit
8、y among the global workforce.As such,integrating AI into our workplaces is a balancing act between seizing opportunities and managing potential disruptions.This paper is intended to serve as a toolkit for businesses to provide guidance on strategies to maximize the potential of employees as they ada
9、pt,learn and grow with these technologies in their professional careers.This paper also serves as a call to action for business leaders to take the initiative in curating strategies and practices that facilitate LLMs to work for businesses,employees and society as a whole.This paper is a direct foll
10、ow-up on the previous edition of the Jobs of Tomorrow series,Jobs of Tomorrow:Large Language Models and Jobs,which took a structured approach to understanding the direct impact of LLMs on specific jobs.That paper provided a structured analysis of the potential impact of LLMs on jobs,enabling stakeho
11、lders business leaders,policy-makers,workers and the broader public to make more informed decisions.In addition,the structured approach proposed in that paper also provided a case study for future waves of technological advancement across sectors.We are deeply grateful to the Centre for the New Econ
12、omy and Society partners and constituents for their leadership of the jobs agenda,as well as for the partnership of the Accenture team,whose members served as core collaborators on this paper.The insights of this paper will be instrumental for the Jobs Consortium,a global coalition of ministers and
13、chief executive officers that promotes a better future of work,through boosting labour market foresight,driving job creation and improving job quality while enabling job transitions and will serve as a key tool for the Good Work Alliance,a global,cross-industry group of businesses committed to prior
14、itizing good work in an evolving jobs landscape.Mary Kate Morley Ryan Talent and Organization/Human Potential Managing Director,AccentureElselot Hasselaar Head of Mission,Work,Wages and Job Creation,World Economic ForumJobs of Tomorrow Large Language Models and Jobs A Business ToolkitDecember 2023Jo
15、bs of Tomorrow Large Language Models and Jobs A Business Toolkit3Executive summaryThe explosive popularity and accessibility of generative artificial intelligence(AI)technologies,such as GitHubs Copilot,Midjourney and ChatGPT,have produced AIs first true inflexion point in public adoption,demonstrat
16、ing the technologys transformative potential.While there are myriad possible use cases for generative AI,including image,video,written language and music creation,large language models(LLMs)have an outsized potential to impact the greatest number of jobs in the near future.The Future of Jobs Report
17、2023 states that global business leaders believe that 23%of global jobs are expected to transform within the next five years due to rapid technological advances,particularly in generative AI technologies.Especially,LLMs have the potential for both automating and augmenting tasks across various occup
18、ations.The Jobs of Tomorrow:Large Language Models and Jobs white paper,published by the World Economic Forum and Accenture earlier in 2023,found that 40%of working hours could be transformed,affecting roles from specialized fields requiring advanced degrees to those focused on routine procedures.Job
19、s emphasizing face-to-face interaction are likely to be less affected.Businesses are urged to adopt proactive and responsible approaches to manage this transformation,addressing concerns like job change and job displacement risk,job quality and skilling.Structured analysis,planning and proactive pre
20、paration are required of business leaders to ensure that LLMs and other technological advancements lead to an improved future of work with new opportunities for workers.This paper serves as a toolkit for businesses,providing practical strategies to navigate the changing job landscape in three primar
21、y areas of consideration:1)job change and job displacement risk,2)job quality,and 3)learning and skilling.Each of these considerations is addressed through business practices guided by three fundamental approaches:1)promoting worker awareness,2)facilitating organizational change,and 3)shifting workp
22、lace norms and culture.The result is a business strategy matrix for effectively leading the workforce transition through the large-scale deployment of LLMs.Businesses will need to strategically navigate concerns over job displacement risk,job quality and skilling.Jobs of Tomorrow Large Language Mode
23、ls and Jobs A Business Toolkit4IntroductionUp to 40%of working hours could be transformed by LLMs businesses must take the lead.Technology is rapidly transforming the labour market.The Future of Jobs Report 2023 found that global business leaders expect 23%of current jobs to transform within the nex
24、t five years.1 This transformation is driven by key trends like technology adoption,the green transition and the global macroeconomic outlook.In particular,technology adoption is driving change in labour markets:82%of business leaders expect increased adoption of new technologies to drive business t
25、ransformation,while 37%of business leaders anticipate that new technologies will be a net job creator,and 21%believe that they will be a net job displacer.Throughout 2023,specifically,generative artificial intelligence(AI)has developed at a rapid pace in terms of capabilities and adoption,and that p
26、ace shows no signs of slowing down.In fact,three out of every four companies across the globe are expected to adopt technologies that include generative AI in the next 3 to 5 years,and 98%of global executives agree AI foundation models will play an important role in their organizations strategies in
27、 that same time period.2 GitHubs Copilot,Midjourney and ChatGPT are a case in point,demonstrating the technologys transformative potential.One week into its November 2022 launch,ChatGPT,OpenAIs trained language model,reached one million users.By January,ChatGPT had reached one hundred million monthl
28、y active users,making it the fastest-growing consumer application in history at the time.3 Generative transformer models have the ability to impact all classes of creative work,possessing the capability to generate novel images,videos,music,sounds and written language.While there are myriad possible
29、 use cases for generative AI,large language models(LLMs)and their unique language-generating capabilities have an outsized potential to impact the greatest number of jobs in the near future.LLMs,such as those powering ChatGPT,have demonstrated this rapid development and have the potential to impact
30、many jobs and workers across industries.In September 2023,to increase understanding of how LLMs could impact jobs,the World Economic Forum,in collaboration with Accenture,published the Jobs of Tomorrow:Large Language Models and Jobs white paper.It provides a structured analysis of the potential Jobs
31、 of Tomorrow Large Language Models and Jobs A Business Toolkit5impact of LLMs on jobs and finds that it will both automate and augment job tasks,ultimately having a transformative effect on jobs.The September 2023 white paper found that an estimated 62%of total work time across occupations involves
32、language-based tasks tasks that could be exposed to the potential impacts of LLMs.Over 19,000 individual tasks across 867 occupations were assessed to understand the potential exposure of each task to LLM adoption,classifying them as tasks with high potential for automation(the task could be perform
33、ed by LLMs,without humans),high potential for augmentation(the task will continue to be performed by humans,with LLMs increasing human productivity),low potential for either(humans will continue to perform the task with no significant impact from LLMs)or unaffected(non-language tasks).The analysis f
34、ound that over 40%of working hours could be transformed by LLMs through automation or augmentation.Jobs with the highest potential for automation of tasks include those that emphasize routine and repetitive procedures and do not require a high degree of interpersonal communication,including many cle
35、rical occupations.Jobs with the highest potential for augmentation by LLMs are those that emphasize critical thinking and complex problem-solving skills,especially those in science,technology,engineering and mathematics(STEM)fields.They include a mix of highly specialized roles often requiring advan
36、ced degrees,such as computer programmers,plus roles that require human validation,such as assessors,monitors and screeners.Jobs emphasizing face-to-face communication and interpersonal interactions are expected to be less exposed to the impacts of LLMs,and those emphasizing non-language tasks will b
37、e even less exposed or not exposed at all.4 Lastly,the paper also found that new jobs could emerge,for example,AI model and prompt engineers,interface and interaction designers,AI content creators,data curators and trainers,and ethics and governance specialists.Given the likely ubiquitous implementa
38、tion of LLMs in business environments,paired with its potential for job transformation as found by the Jobs of Tomorrow:Large Language Models and Jobs white paper,it is critical that businesses adopt a proactive,human-centred and responsible approach to LLMs.This white paper is a follow-up to the Jo
39、bs of Tomorrow:Large Language Models and Jobs paper and serves as a toolkit for businesses to address three considerations for the technological transformation of labour markets to come.The first is job change and job displacement risk:as LLMs are capable of performing many of the language tasks use
40、d on the job,and up to 60%of work time uses language tasks,there is potential for some jobs to be displaced as others emerge.The second is job quality:the deployment of these technological tools could affect not only the number of jobs but also the nature of the work itself,potentially affecting the
41、 meaningfulness of work,the well-being of workers and the diversity associated with new and legacy roles.The third concern is learning and skilling:to adjust to and embrace technological change requires continuous learning;an agile workforce is the best approach to meet the evolving demands of the l
42、abour market.In the following sections,each consideration is first defined and then addressed with several practical business strategies that ensure a cohesive and inclusive approach to the rapidly changing job landscape.This white paper proposes that the business strategies fall into three primary
43、approaches:1)promoting worker awareness,which involves informing and educating employees about the potential impacts of LLMs on their current roles and future job prospects,empowering them to proactively engage with the evolving technological landscape,2)facilitating organizational change,in which c
44、ompanies adapt their structures,processes and strategies to harness the benefits of new technologies while mitigating the risks associated with such transformations,and 3)shifting workplace norms and culture,to reshape the underlying attitudes,behaviours and values within the workplace to promote ag
45、ility in responsiveness to change.Table 1 summarizes the key considerations and proposed business strategies.Summary of business strategies in each of the primary three concernsTABLE 1Job change and job displacement riskJob qualitySkilling and learningWorker awareness1 Deploy proactive forecasting a
46、nd communicate expectations4 Ensure inclusive design of AI and LLM applications7 Roll out broad LLM literacy programmes alongside AI-specific jobsOrganizational change2 Set up and make use of internal job markets5 Implement transparent governance structures for LLM deployment8 Apply a skill-first ap
47、proach organization wideShifting norms and culture3 Create increased workforce agility6 Increase awareness of LLM benefits in the workplace9 Offer workplace learning opportunities Over 40%of working hours could be transformed by LLMs through automation or augmentation.Jobs of Tomorrow Large Language
48、 Models and Jobs A Business Toolkit6Consideration 1 Job change and job displacement risk1Proactively addressing job change and job displacement risk through internal job markets and worker agility ensures smooth transitions for employees.Concerns have arisen that the latest wave of generative AI tec
49、hnologies,specifically LLMs,could lead to the automation of job tasks and,ultimately,job loss.While labour market predictions are never 100%certain,recent World Economic Forum research took a structured approach and found that the adoption of LLMs could automate some tasks and augment others.5 The W
50、orld Economic Forums Jobs of Tomorrow:Large Language Models and Jobs white paper found that job tasks with the highest potential for automation by LLMs tend to be those that use language in a routine and repetitive manner.While LLMs will impact tasks rather than jobs as a whole jobs emphasizing thes
51、e types of language tasks could see a decline,with some workers aware of this potential.For example,over the course of Q3 of 2023,global searches for“Is my job safe?”doubled,indicating the publics increasing concern about whether LLMs would automate their jobs.6 However,LLMs and other generative AI
52、technologies offer potential for job growth as well.The same paper found that job tasks emphasizing abstract reasoning and problem-solving skills have the highest potential for augmentation by LLMs,meaning these technologies will assist in task execution and boost individual productivity.Additionall
53、y,the paper found there is potential to generate new employment opportunities by creating roles that oversee,complement and enhance AIs capabilities.The Future of Jobs Report 2023 echoes many of these findings,indicating that many of the jobs with high potential for automation by LLMs were also expe
54、cted by business leaders to undergo employment decline within the next five years,such as bank tellers and related clerks,data entry clerks,and administrative and executive secretaries.Jobs with higher potential for augmentation are expected to grow,such as AI and machine learning specialists,data a
55、nalysts and scientists,and database and network professionals.Nevertheless,labour market predictions are never completely certain and will likely evolve with the development of LLMs.Businesses must be proactively engaged in addressing the concerns of their workers while preparing them to adapt to th
56、e transition to come.Job tasks emphasizing abstract reasoning and problem-solving skills have the highest potential for augmentation by LLMs,meaning these technologies will assist in task execution and boost individual productivity.Worker awareness deploy proactive forecasting and communicate expect
57、ationsUsing predictive analytics to forecast which jobs will most likely be affected by LLMs and other disruptions,and communicating actively with management and employees to proactively provide pathways to mitigate negative employment impacts.Organizations have the ability and responsibility to sup
58、port workers through job changes and disruptions.Equipping employees with the right knowledge empowers them to make better decisions7 in searching for reskilling opportunities or transitioning to a different job or career.Clearly sharing data and methods will also increase trust and improve employee
59、 well-being,performance and business accountability.8 Analyses like Jobs of Tomorrow:Large Language Models and Jobs provide realistic assessments of the labour market and can help shape expectations in the short term.Organizations can use the results of these studies or apply similar analysis method
60、s to understand how LLMs could transform their business and the work of their employees.In 2019,Singapore launched its National AI Strategy,which includes additional skilling and education programmes to support workers particularly relevant as a 2018 report by Cisco and Oxford Economics found that S
61、ingapore had the most relative exposure of six ASEAN countries to potential job displacement effects of AI,with 7 in 10 workers in Singapore saying AI will impact their jobs.9 Furthering this commitment to skilling and workforce resilience,the Infocom Media Development Authority(IMDA),for example,re
62、cently committed to reskilling 18,000 Singaporeans in AI and adjacent skills through these skilling programmes.10 These efforts have borne fruit since the introduction of the programme in 2015,1 in 4 eligible Singaporeans has participated in SkillsFuture Singapore across many industries,participatin
63、g companies and education programmes.11Organizational change set up and make use of internal job marketsCreate a dynamic internal job marketplace that helps employees transition into roles as technology and business conditions change.Developing internal job markets enables employees to understand th
64、e opportunities available to them within the firm and provides a chance to grow.Enabling internal mobility not only rewards loyal employees,it enables firms to address their talent shortages while mitigating costs.About 60%of companies globally note that skills gaps are a major limitation to busines
65、s transformation.12 Replacing workers is costly:Gallup estimates that the replacement cost can exceed two times an employees annual salary.13 Retaining experienced workers through job transitions on the internal job market reduces total hiring costs and increases the speed at which people contribute
66、 value to an organization.HSBC,a global financial services organization based in the UK,deployed an internal job market for about 200,000 of its employees with plans to expand access to all 220,000 employees around the world.They estimate that their internal talent marketplace is central to their sk
67、ill-building strategy and has unlocked over 170,000 hours of work.Furthermore,about 45%of projects are cross-functional.14Shifting norms and culture create increased workforce agilityCreate a company culture that values and rewards flexibility in job roles,encouraging employees to embrace diverse wo
68、rk experiences and interdisciplinary skill development,which could reduce the anxiety around job displacement and promote a resilient workforce.By cultivating a culture that values workforce agility,employees become more adaptable and resilient,better equipped to transition into new roles as LLMs im
69、pact job markets.Creating such a company culture involves a deliberate approach that starts with leadership commitment and the clear articulation of a new vision and values.It involves engaging employees,aligning hiring and reward systems to these values,and reinforcing them through consistent polic
70、ies and communication.A good example is“side project time”,or“the 20%concept”,popularized by Google but which is also currently used by firms like Atlassian15 and Rakuten.16 These programmes have enabled employees to focus on personal projects not directly tied to a job task during their time at wor
71、k.Side projects enable employees to explore personal interests and build transferable skills outside their job specialization,helping retention,improving morale and encouraging creativity.17 It also enables companies to boost their innovative potential something large companies struggle with.18 The
72、output from these programmes can be significant:Gmail and AdSense both arose from past side projects at Google.191.1 Responsible business strategiesJobs of Tomorrow Large Language Models and Jobs A Business Toolkit8Consideration 2 Job quality 2Maintaining high job quality involves worker inclusion i
73、n LLM governance,involving diverse teams in LLM development and promoting awareness of LLM benefits.Concerns have also arisen that LLMs and other generative AI technologies could erode job quality,potentially making jobs and workplaces less fulfilling and less inclusive and fair.Additionally,increas
74、ing concerns about bias in LLMs highlight the risk of inherent systemic discrimination in these models,which could exacerbate existing inequalities.Despite these concerns,emerging data show that LLMS and AI technologies also have the potential to improve job quality.As discussed in Jobs of Tomorrow:
75、Large Language Models and Jobs,the use of LLMs could automate the more routine and repetitive tasks on the job,which could otherwise disengage workers,while making room for more creative,problem-solving and independent decision-making tasks on the job,boosting worker fulfilment.The World Economic Fo
76、rums Good Work Framework offers a robust definition of job quality and a strategic template to maintain and improve the quality of work in rapidly evolving labour markets.It outlines key objectives to create a better future of work,including promoting fair wages and responsible technology use,provid
77、ing flexibility and protection for all workers,ensuring health and well-being,driving diversity,equity and inclusion,and creating a culture of continuous learning and employability.Of particular relevance when incorporating LLMs and other AI technologies on the job is the frameworks objective of mai
78、ntaining the health and total well-being of workers.There is not yet strong evidence that LLMs have impacted workers well-being.20 Yet,worker anxiety has increased.A 2023 survey by the American Psychological Association found that 38%of workers are worried about the impact of AI on their jobs;roughl
79、y half of those with these concerns say that work negatively affects their mental health,compared to only 29%of those without concerns.21 Another critical concern is that LLMs could induce bias in responses that lead to decision-making,highlighting the importance of adhering to the Good Work Framewo
80、rks objective of driving diversity,equity and inclusion.For instance,persistent patterns of racial and ethnic discrimination could be reflected in hiring datasets,which could skew algorithms used to sort and filter job candidates in the application screening process.22 In the same vein,most LLMs are
81、 trained on English texts,which could lead LLMs to produce overwhelmingly Anglo-centric responses.23 Furthermore,the burdens of adapting LLM-induced changes in jobs could disproportionately fall on particular population groups who may be over or underrepresented in affected occupations,such as women
82、 and ethnic minorities.24 Businesses must move proactively and with urgency to address these concerns,particularly around job quality,agency and inclusivity,if they are to establish the trust needed to unlock the positive potential of LLMs.LLMs could automate the more routine and repetitive tasks on
83、 the job,which could otherwise disengage workers,while making room for more creative,problem solving and independent decision-making tasks on the job,boosting worker fulfilment.Jobs of Tomorrow Large Language Models and Jobs A Business Toolkit92.1 Responsible business strategiesWorker awareness ensu
84、re inclusive design of AI and LLMs applicationsEmphasize the collaborative and inclusive development of AI applications and LLMs.This involves establishing cross-functional teams with diverse backgrounds and actively involving employees in the design and implementation of company-specific AI systems
85、 and LLMs.LLMs have the greatest positive impact when developed,tested and piloted by teams with a diverse mix of skills and perspectives.Driving diversity among the people who build LLM applications can help reduce the built-in bias those applications and algorithms demonstrate.Diverse,representati
86、ve teams can think through the operational changes new LLM applications require,and development teams that include end users result in applications with higher chances of adoption.In total,92%of companies that scale AI effectively use multidisciplinary teams.25 Microsoft,for example,has created a fr
87、amework for assembling“inclusive”technology design teams to support inclusion,fairness,reliability and safety.26 Paired with inclusive skilling initiatives,efforts such as this can drive improved workforce representation and reduce bias.Similarly,Siemens actively involves its engineering and shop fl
88、oor employees in developing AI applications for industrial use.This involvement ensures that the AI solutions are practical and enhance operational efficiency in real-world industrial settings.27Organizational change implement inclusive governance structures for LLM deploymentEnsure the adoption of
89、transparent and inclusive LLM governance and frameworks.This approach necessitates inclusive and participatory decision-making processes in all aspects of LLMs governance and deployment,addressing not only efficacy but also bias,ethical implications and overall impact on job quality.Transparency aro
90、und the deployment of LLMs,detailing their use,is key to building employee trust.When employees fully understand how AI,and specifically LLMs,apply to their role,the adoption of the technology is two times more likely to be successful.28 This transparency is best achieved through establishing respon
91、sible LLM frameworks that involve employees throughout.The framework should,for example,detail how a participative decision culture should guide the design,development and deployment of LLMs.29 A vital element of these frameworks is the inclusion of employee feedback mechanisms akin to those propose
92、d in the World Economic Forums Good Work Framework.30 Such mechanisms ensure ongoing assessment and refinement of LLM applications,focusing on their ethical,impartial and effective use within the company.Several leading companies have been recognized for their efforts in adopting transparent and inc
93、lusive AI governance,which can be tailored to LLMs specifically.Google has implemented a three-tiered governance structure to oversee the responsible use of AI.This structure includes multidisciplinary reviewers and product team members whose primary focus is on developing AI applications that yield
94、 social benefits.31 IBM is known for its commitment to trustworthy AI;they have an AI Ethics Board and have published clear principles on trust and transparency in AI.IBM emphasizes AI that is explainable,fair and secure,and they provide tools for businesses to detect and mitigate bias in AI systems
95、.32 Accenture established a responsible AI(RAI)compliance programme to institutionalize its RAI principles into practice by embedding RAI across its entire enterprise from development and use,to deployment and post-deployment monitoring.Shifting norms and culture increase awareness of LLM benefits i
96、n the workplaceDrive the adoption of LLMs by encouraging innovation and increasing awareness of how LLMs can benefit workers,including by automating routine tasks and augmenting human potential,resulting in more meaningful and fulfilling jobs.As discussed in Jobs of Tomorrow:Large Language Models an
97、d Jobs,LLMs could significantly improve the overall work experience by automating routine tasks and augmenting others,resulting in more meaningful and fulfilling jobs and freeing time and resources to spur innovation.Roughly 60%of organizations are optimistic about the impact that generative AI will
98、 have on their people,including their overall work experience.Approaches include improving job satisfaction by reducing time spent on routine tasks and enabling individuals to engage in more meaningful and innovative tasks.33 Effectively communicating these positive impacts and examples is key to ad
99、dressing concerns around job quality.An example case study comes from Moderna,whose employees are frequent users of AI technologies,including LLM tools.Moderna used AI tools to develop mRNA COVID-19 vaccines and continues to find new uses for them.One executive noted that removing repetitive,non-val
100、ue-add tasks from employees responsibilities boosts morale at Moderna.Employees are also eager for this change,as they recognize the benefits of engaging in insightful and creative work rather than monotonous tasks.34 Jobs of Tomorrow Large Language Models and Jobs A Business Toolkit10Consideration
101、3 Skilling and learning3Enhancing workforce skills with LLM training and a skills-first approach ensures employees are empowered to leverage LLMs optimally.With the ascent of LLMs and other AI systems,it has become imperative for organizations to embrace a culture of continuous learning and skill de
102、velopment.Additionally,many workers have doubts surrounding their current skill levels.The Accenture Future of Work Study 2022 showed that only 31%of workers are confident that their skills will be relevant in five years.35 Over 50%of employees believe LLM skills specifically will be essential to th
103、eir role,according to a global study done by the staffing agency Randstad.36Employers have also expressed similar concerns:according to the Future of Jobs Report 2023,business leaders estimate that 44%of workers skills will be disrupted in the next five years.Businesses estimate that six in 10 worke
104、rs will require training before 2027,yet only half of workers are seen to have access to adequate training opportunities today.37 Training will naturally need to focus on the skills predicted to grow the most,namely analytical skills,creative thinking,technology literacy,curiosity and lifelong learn
105、ing,resiliency,flexibility and agility,which three out of four firms predict will rise in importance.AI and big data skills rank as the number three priority in company training strategies between 2023 and 2027 and are the number one priority for companies with over 50,000 employees.The need for AI
106、skills is also evident directly in employment data.Employer demand for LLM and other generative AI skills is growing.AI skills are rare and in high demand,creating a new marketplace for highly paid talent while widening reskilling opportunities across adjacent skillsets.38 Across 15 advanced economi
107、es,AI hiring is up by about 60%since 2016.In an analysis of job postings in 2022,jobs requiring AI skills reached 2.05%of postings in the United States.39 By focusing on the development of skills,from core digital literacy to deep AI expertise,organizations have an opportunity to fill critical talen
108、t gaps while improving the overall working experience of their people.Business leaders estimate that 44%of workers skills will be disrupted in the next five years.11Jobs of Tomorrow Large Language Models and Jobs A Business ToolkitWorker awareness roll out broad LLM literacy programmes along AI-spec
109、ific jobsIntroduce courses that explain LLMs,including their limitations and how they can be used,to demystify the technology for workers.By establishing a baseline level of LLM fluency among their entire workforce,companies can help their employees understand what LLMs are,how their jobs will evolv
110、e to work alongside the technology,and its benefit to the enterprise when responsibly deployed.This approach not only builds confidence in LLMs but also encourages adoption and use.Employees should be trained to use LLMs to augment tasks where applicable,especially for in-demand occupations and crit
111、ical processes.This will also support the development of the workforces needed to fill current and emerging AI-related jobs.Examples include Microsofts new AI Skills initiatives,which include free certifiable coursework to improve data fluency,40 and courses like DeepLearning.AIs Generative AI for E
112、veryone on Coursera,which provides accessible,comprehensive and practical coverage of the technology.Employees engaged in AI deployment,such as Data Scientists and Machine Learning Engineers,can benefit from more advanced training programmes,such as Amazons library of AI and machine learning modules
113、,which include over 80 online courses.41Organizational change apply a skills-first approach organization-wideFacilitate employee transitioning and hiring,as well as upskilling,by assessing skills directly,according to jobs skills requirements.A skills-first approach focuses on whether a person has t
114、he right skills and competencies for a particular role,rather than having the right degree,job history or previous job titles,helping to ensure that businesses get the skills they need for a particular job while democratizing access to jobs for those who have the competencies but not the right forma
115、l qualifications for a role.42 The World Economic Forums Putting Skills First:A Framework for Action paper provides a comprehensive framework for action that businesses and governments can use to shape their skills-first approach.Some studies demonstrate that companies hiring based on skills are 36%
116、less likely to face talent and skills shortages.43 An example is Henkel,a German-headquartered consumer and industrial company,which developed a digital talent-skilling programme for about 10,500 of its managers.The programme involved assessing digital skills capabilities,identifying skills gaps,ups
117、killing,and developing a Recruiter Toolbox to ensure that digital skills were included in the sourcing of new hires.Recruiting was also streamlined,as internal employees seeking new roles could easily see the skills requirements for a role,and then take the relevant courses to gain the required skil
118、ls,becoming viable candidates.44Shifting norms and culture offer workplace learning opportunitiesOffer work-based learning opportunities such as apprenticeships,rotational programmes and temporary assignments to keep employees skills current and to encourage a culture of lifelong learning.One of the
119、 most effective ways to upskill and reskill workers is by offering them on-the-job learning experiences.Apprenticeships provide long-term,paid,work-based learning opportunities alongside structured curricula that enable learners to gain both education and hands-on skills related to an occupation.45
120、Rotational programmes are structured programmes offered by employers that allow employees,typically recent graduates or new hires,to gain experience in different areas or departments of a company over a set period of time.Employee retention rates are higher for companies that offer rotational progra
121、mmes.46 Temporary work assignments are short-term positions or projects that employees take on for a defined period.These are beneficial as they provide flexibility and diversity of experience to the workforce.Microsoft,for example,launched its Leap programme to facilitate recruitment in tech.Micros
122、oft Leap is a 16-week immersive traineeship for individuals with baseline technical expertise,which combines classroom learning with hands-on engineering projects on existing products such as Azure,Xbox,Bing and Office365.As a result,98%of Leap participants ultimately find work in the tech industry.
123、47 Additionally,Accenture launched its own apprenticeship programme in 2016,alongside the launch of their Apprentice Network,which has since expanded to 10 locations with 195 unique employers across North America.Accenture has set a goal for apprentices to make up 20%of North America entry-level hir
124、ing.48 3.1 Responsible business strategiesJobs of Tomorrow Large Language Models and Jobs A Business Toolkit12ConclusionThe revolutionary impact of generative AI and,more specifically,LLMs is undeniable,and organizations globally are seeking ways to capitalize on its transformative capabilities.This
125、 transformation must be managed responsibly.To fully benefit from AI and LLMs,businesses must first address concerns around ethics,trust,governance and legality.Yet,today,only 6%of organizations have adopted responsible AI frameworks and put principles into practice49 making the considerations outli
126、ned in this paper all the more relevant.As organizations address the risks presented by LLMs including fairness,bias,untrustworthy outputs and legal considerations they must also understand and address the impact the technology will have on their workers.In this nascent stage of the next wave of tec
127、hnological evolution,businesses have an opportunity to pioneer effective strategies for integrating LLMs in a manner that benefits workers and broader society.This paper serves as a toolkit for businesses to address three key areas for consideration in implementing generative AI and LLMs in the work
128、place.The first is job change and job displacement risk:businesses must proactively manage the risk of job displacement by implementing transparent forecasting,establishing dynamic internal job markets and encouraging workforce agility.The second is job quality:businesses must uphold high standards
129、for job quality by embracing inclusive design principles,establishing collaborative governance structures and nurturing a work culture built around innovation and awareness.The third and final consideration is skilling:businesses will need to keep pace with the evolving needs for workforce capabilit
130、ies by focusing on implementing AI literacy programmes,adopting a skills-first approach to talent management and championing learning on the job.The insights of this paper will be instrumental in guiding the progress of the World Economic Forums Jobs Consortium,a global coalition of ministers and ch
131、ief executive officers that promotes a better future of work through boosting labour market foresight,driving job creation and improving job quality while enabling job transitions.It will also serve as a key tool for the Good Work Alliance,a global,cross-industry group of businesses committed to pri
132、oritizing good work in an evolving jobs landscape.Business leaders must take a structured and proactive approach to integrating LLMs,ensuring that these technologies benefit businesses,employees and society at large.By adopting the strategies outlined in this paper,businesses can ensure a positive w
133、orkforce transformation,harnessing the potential of LLMs to drive innovation,productivity and job satisfaction.Jobs of Tomorrow Large Language Models and Jobs A Business Toolkit13ContributorsAcknowledgementsWorld Economic Forum Elselot HasselaarHead,Mission for Work,Wages and Job CreationAndrew Silv
134、aInsights Lead,Work,Wages and Job CreationSaadia ZahidiManaging DirectorAccentureNathan DecetyManager,Accenture StrategyPaul Daugherty Chief Technology and Innovation Officer Hernan EspinosaSenior Manager,Accenture StrategyAllison HornTalent and Organization/Human Potential Managing Director Mary Ka
135、te Morley RyanTalent and Organization/Human Potential Managing DirectorChristine NananPrincipal Director,Accenture StrategyKathleen OReillyGlobal Lead,Accenture Communications,Media and Technology Leila YosefSenior Principal,Accenture ResearchWe are further grateful to our colleagues at the Centre f
136、or the New Economy and Society for helpful suggestions and comments,in particular the authors of the Future of Jobs Report 2023,Sam Grayling,Ricky Li,and Mark Rayner,and Manager of the Good Work Alliance,Isabelle Leliaert.We extend thanks to many Accenture colleagues for providing insights and guida
137、nce that assisted in the shaping of this paper,including Colin Anderson,Gabriela Burlacu,Arnab Chakraborty,Liz Creamer,Lan Guan,Laurie Henneborn,Andrew Levy,Ali Shah,Kathryn White and Jim Wilson.For more information,or to get involved,please contact cnesweforum.org.ProductionLaurence Denmark Creativ
138、e Director,Studio MikoMartha Howlett Editor,Studio MikoJobs of Tomorrow Large Language Models and Jobs A Business Toolkit14Endnotes1.World Economic Forum,Future of Jobs Report 2023,2023.2.Accenture,Technology Vision 2023,2023.3.Mollman,Steve,“ChatGPT gained 1 million users in under a week.Heres why
139、the AI chatbot is primed to disrupt search as we know it”,Yahoo Finance,9 December 2022,https:/ sets record for fastest-growing user base analyst note”,Reuters,2 February 2023,https:/ Economic Forum,Jobs of Tomorrow:Large Language Models and Jobs,2023;Eloundou,Tyna,Sam Manning,Pamela Mishkin and Dan
140、iel Rock,“GPTs are GPTs:An Early Look at the Labor Market Impact Potential of Large Language Models”,Cornell University,17 March 2023,https:/arxiv.org/abs/2303.10130;World Economic Forum,Future of Jobs Report 2023,2023.5.World Economic Forum,Jobs of Tomorrow:Large Language Models and Jobs,2023.6.“AI
141、 is not yet killing jobs”,The Economist,15 June 2023,https:/ Advantages of Data-Driven Decision-Making”,Harvard Business School,26 August 2019,https:/online.hbs.edu/blog/post/data-driven-decision-making.8.Maloney,Devon,“Transparency in business:the next wave in company evolution”,Slack,30 November 2
142、019,https:/ to Build Trust in the Workplace”,Gallup,14 June 2022,https:/ Business Review,The Impact of Employee Engagement on Performance,2013,https:/hbr.org/resources/pdfs/comm/achievers/hbr_achievers_report_sep13.pdf.9.Cisco,Technology and the future of ASEAN jobs,2018,https:/ Media Development Au
143、thority(IMDA),IMDA leads nationwide AI skilling to build AI talent pool Press release,4 September 2023,https:/www.imda.gov.sg/resources/press-releases-factsheets-and-speeches/press-releases/2023/imda-leads-ai-skilling-to-build-ai-talent-pool.11.SkillsFuture Singapore,Strong Participation in Skillsfu
144、ture Programmes;SSG Seeks to do More with Enterprises Press release,23 March 2023,https:/www.skillsfuture.gov.sg/newsroom/strong-participation-in-skillsfuture-programmes-ssg-seeks-to-do-more-with-enterprises#:text=23%2F03%2F2023-,STRONG%20PARTICIPATION%20IN%20SKILLSFUTURE%20PROGRAMMES%3B%20SSG%20SEE
145、KS%20TO%20DO%20MORE,Singapore%20(SSG)%2Dsupported%20programmes.12.World Economic Forum,Future of Jobs Report 2023,2023.13.McFeely,Shane and Ben Wigert,“This Fixable Problem Costs U.S.Businesses$1 Trillion”,13 March 2019,https:/ Upskilling and Reskilling Opportunities for our Employees”,World Economi
146、c Forum,https:/initiatives.weforum.org/good-work-alliance/case-studies/hsbc:-providing-upskilling-and-reskilling-opportunities-for-our-employees/aJY68000000PAvbGAG#.15.Rotenstein,John,“Atlassians 20%Time:A Year in Review”,Atlassian,19 February 2009,https:/ Culture”,Rakuten,n.d.,https:/ Positive Effe
147、ct of Creative Hobbies on Performance at Work”,PsyBlog,28 April 2014,https:/www.spring.org.uk/2014/04/the-positive-effect-of-creative-hobbies-on-performance-at-work.php?ref=.18.Wessel,Maxwell,“Why Big Companies Cant Innovate”,Harvard Business News,27 September 2012,https:/hbr.org/2012/09/why-big-com
148、panies-cant-innovate;Viki,Tendayi,“Why Large Companies Continue To Struggle With Innovation”,Forbes,4 November 2018,https:/ 20%rule shows exactly how much time you should spend learning new skillsand why it works”,CBNC,16 December 2021,https:/ Knig and Luca Stella,Artificial Intelligence and Workers
149、 Well-Being,IZA Institute of Labor Economics,2023,https:/docs.iza.org/dp16485.pdf.21.American Psychological Association,2023 Work in America Survey,2023,https:/www.apa.org/pubs/reports/work-in-america/2023-work-america-ai-monitoring.22.Quillan,Lincoln and John J.Lee,“Trends in racial and ethnic disc
150、rimination in hiring in six Western countries”,PNAS,31 January 2023,https:/www.pnas.org/doi/10.1073/pnas.2212875120.Jobs of Tomorrow Large Language Models and Jobs A Business Toolkit1523.Nicholas,Gabriel and Aliya Bhatia,Lost in Translation:Large Language Models in Non-English Content Analysis,Cente
151、r for Democracy&Technology(CDT),2023,https:/arxiv.org/ftp/arxiv/papers/2306/2306.07377.pdf.24.Gmyrek,Pawe,Janine Berg,David Bescond,Generative AI and Jobs:A Global Analysis of Potential Effects on Job Quality and Quantity,ILO,2023,p.39,https:/www.ilo.org/wcmsp5/groups/public/-dgreports/-inst/documen
152、ts/publication/wcms_890761.pdf.25.Lyman,Michael,“Scaling AI:Giving data its due”,Accenture,9 November 2021,https:/ R.,H.James Wilson and Rumman Chowdhury,“Using Artificial Intelligence to Promote Diversity”,MIT Sloan Management Review,21 November 2018,https:/sloanreview.mit.edu/article/using-artific
153、ial-intelligence-to-promote-diversity/.27.Siemens,Siemens and Microsoft drive industrial productivity with generative artificial intelligence Press release,12 April 2023,https:/ to Scale,2019,https:/ AI results by“going pro”,Accenture,24 September 2020,https:/ Economic Forum,The Good Work Framework:
154、A new business agenda for the future of work,2022.31.Monetary Authority of Singapore,FEAT Principles Assessment Case Studies,n.d.,https:/www.mas.gov.sg/-/media/mas/news/media-releases/veritas-document-6-feat-principles-assessment-case-studies.pdf.32.“AI Ethics”,IBM,n.d.,https:/ Pulse of Change”,Acce
155、nture,September 2023,https:/ expect AI to improve jobs,productivity”,Tech Target,22 June 2023,https:/ Cultural Benefits of Artificial Intelligence in the Enterprise”,MITSloan,2 November 2021,https:/sloanreview.mit.edu/projects/the-cultural-benefits-of-artificial-intelligence-in-the-enterprise/.35.Ac
156、centure,Accenture Future of Work Study 2022,2022,https:/ than half of workers believe AI skills will future-proof their careers,but only 13%have been offered such training opportunities-Randstad data reveals Press release,5 September 2023,https:/ Economic Forum,Future of Jobs Report 2023,2023.38.“Ar
157、tificial Intelligence skills can increase salaries by as much as 40%”,University of Oxford,24 October 2023,https:/www.ox.ac.uk/news/2023-10-24-artificial-intelligence-skills-can-increase-salaries-much-40#:text=Workers%20with%20artificial%20intelligence%20skills,Institute%2C%20and%20the%20Center%20fo
158、r.39.Stanford University,Artificial Intelligence Index Report 2023,2023.40.“New skills for a new way of working”,Microsoft,n.d.,https:/ Skill Builder”,Amazon,n.d.,https:/skillbuilder.aws/.42.World Economic Forum,Putting Skills First:A Framework for Action,2023,https:/www3.weforum.org/docs/WEF_CNES_P
159、utting_Skills_First_2023.pdf.43.Accenture,Uncover Missed Talent Pools,Improve Diversity,2021,https:/ a global workforce to be future-ready”,Accenture,n.d.,https:/ three high schoolers have to say on the transformative impact of youth apprecnticeships”,27 March 2023,https:/www.brookings.edu/articles/
160、what-three-high-schoolers-have-to-say-on-the-transformative-impact-of-youth-apprenticeships/.46.Glaveli,Niki and Eleonora Karassavidou,“Exploring a Possible Route Through Which Training Affects Organizational Performance:The Case of a Greek Bank”,The International Journal of Human Resource Managemen
161、t 22,no.14,2011,pp.2892-2923,https:/ Lead Apprenticeship Program,n.d.,https:/ art of AI maturity,2022,https:/ of Tomorrow Large Language Models and Jobs A Business Toolkit16World Economic Forum9193 route de la CapiteCH-1223 Cologny/GenevaSwitzerland Tel.:+41(0)22 869 1212Fax:+41(0)22 786 2744contactweforum.orgwww.weforum.orgThe World Economic Forum,committed to improving the state of the world,is the International Organization for Public-Private Cooperation.The Forum engages the foremost political,business and other leaders of society to shape global,regional and industry agendas.