1、04Whos afraid of GenAI?08AI and the insurance industry16 July 2024Allianz ResearchGenAI in the insurance AllianzTradeindustry:Divine coincidence 12The way forwardfor human capital2 Whos afraid of GenAI?While experts predict substantial positive economic effects from the rise of generative artificial

2、 intelligence,public sentiment is not so optimistic.In our survey of over 6,000 people in Austria,France,Germany,Italy,Poland and Spain,36%expressed concerns over the risks presented by AI,with 46%expecting AI to cut the number of jobs available(vs.33%who expect AI to increase the number of jobs ava

3、ilable).More worryingly,more than half of all respondents(51%)believed that the skills gap and inequality could widen as AI uptake expands across industries,with the smart getting smarter and the rest being left behind.Just 21%of all respondents were optimistic about the benefits of AI for their eco

4、nomies.But we find that fears of AI causing a massive labor dislocation in the insurance industry are overblown.As a data-driven industry like few others,the insurance industry in particular has significant potential for automation and productivity enhancement.AI applications can complement and augm

5、ent employees skills and are likely to improve efficiency,customer satisfaction and fraud detection.This could lead to labor reductions but economic models suggest only a modest correlation between productivity gains and labor reductions in the insurance sector:a 0.622%increase in productivity would

6、 yield only a 1%decrease in labor.In fact,the timing of the advance of AI seems to be fortunate.It coincides with demographic shifts,i.e.aging populations and shrinking labor forces,which could lead to labor shortages in many sectors,including insurance.By increasing productivity and automating rout

7、ine tasks,AI could help the industry combat this looming challenge.As AI adoption increases,balancing innovation with regulation will be crucial.While many advocate for strict AI regulation to prevent harm,others emphasize the need to maintain competitiveness.Policymakers and industry leaders must n

8、avigate these challenges to harness AIs potential while addressing societal concerns.ExecutiveAllianz ResearchSummary Arne HolzhausenHead of Insurance,Wealth&ESG RPatricia Pelayo-RomeroSenior Economist,Insurance&ESGpatricia.pelayo-16 July 20243 3Whos afraid of GenAI?4Allianz ResearchOver the last de

9、cade,the way we work has undergone a deep transformation,from the focus on diversity,equity and inclusion(DEI)to the shift from traditional career stereotypes to the pioritization of life-long learning and reskilling,to name a few.The pandemic also kicked off a remote-work revolution that has proved

10、 almost impossible to reverse.Now,with the promise of increased productivity brought on by generative artificial intelligence(GenAI),the burning question is,what will be the future of the labor market?Experts believe GenAI will have truly transformative implications for our economies,boosting growth

11、 substantially.For example,Goldman Sachs expects that the adoption of GenAI will boost global GDP by+7%(around EUR7trn)and lead to an annual productivity gain of+1.5%in the US over the next decade.McKinsey predicts that GenAI will produce an even more substantive boost of EUR16trn-EUR 24trn for the

12、global economy.Yet,the public remains much less optimistic.Earlier this year,we asked over 6,000 people in six European countries(Austria,France,Germany,Italy,Poland and Spain)about the potential impact of AI on their economies,jobs and wages.2 Most of them(36%)believed that AI would bring about mor

13、e risks than opportunities,while 29%believed the risks and opportunities would balance each other out.Just 21%were more optimistic,believing that AI would bring about benefits for their economies.The optimism of our respondents was positively correlated with their education level,meaning that the hi

14、gher their reported education level was,the more likely they were to believe in the positive developments AI would bring to their economies.But even among the well-educated respondents,skepticism prevails(Figure 1).Acemoglu,D.(2024).The Simple Macroeconomics of AI.MIT.Link:https:/economics.mit.edu/s

15、ites/default/files/2024-04/The%20Simple%20Macroeconomics%20of%20AI.pdf Allianz Pulse 2024:What unites and separates the demos of Europe16 July 20245 5Figure 1:Whos afraid of AI?In your opinion,how will AI affect the economy of your country?Will AI awake more chances or more risks in the economy?Sour

16、ce:Allianz ResearchRespondents were also concerned about the impact of AI on jobs.The IMF expects 60%of jobs to be affected by the GenAI revolution.As AI improves productivity,some of those jobs might be replaced,but new roles will also be created.In fact,data from LinkedIn show that the share of jo

17、b descriptions mentioning the use of AI has increased more than two-fold in Austria,France,Germany,Italy and Spain since the introduction of ChatGPT in November 2022.At the same time,investment into GenAI tools has skyrocketed:global venture capital investment in the first quarter of 2024 amounted t

18、o EUR2.7bn,while the total investment in GenAI for 2023 was around EUR20bn,driven mainly by two tech giants which accounted for three quarters of all investments among them.In general,respondents seem to believe that AI will lead to fewer jobs:46%of our respondents said AI will make workers more eff

19、icient,allowing companies to reduce jobs.In comparison,35%believed that AI would create new demands,tasks and roles.Interestingly,respondents from the lower-income category were more likely to believe that AI will reduce jobs,with only 27%in this group expecting AI to create new jobs.In contrast,42%

20、of those in the highest income quintile(quintile 5)expect AI to create new jobs(Figure 2).12%19%28%18%12%20%17%23%30%24%21%25%17%22%24%7%16%24%54%46%33%32%40%36%36%37%31%47%34%33%42%31%34%41%39%35%17%27%29%18%26%30%26%28%33%18%29%29%22%31%30%22%31%31%14%7%7%18%18%11%17%9%4%6%14%10%17%13%10%28%11%8%0

21、%10%20%30%40%50%60%70%80%90%100%PrimarySecondaryTertiaryPrimarySecondaryTertiaryPrimarySecondaryTertiaryPrimarySecondaryTertiaryPrimarySecondaryTertiaryPrimarySecondaryTertiaryAustriaFranceGermanyItalyPolandSpainMore opportunitiesMore risksBalance each other outNo influenceI dont knowAllianz Researc

22、h6Source:Allianz ResearchFigure 2:Who can see the bright side of AI?In your opinion,how will AI affect the labor market of your country?Will AI become a job killer or job creator?Share of respondents,in%35%27%35%38%38%42%46%47%47%45%45%43%19%26%18%17%17%16%0%20%40%60%80%100%Totalquintile 1quintile 2

23、quintile 3quintile 4quintile 5I dont knowWith the help of AI,companies become more efficient and can reduce jobsAI will create new demands and tasks and therefore new jobs16 July 20247 7 Brynjolfsson,E.et al.(2023).Generative AI at Work.NBER.Link:https:/danielle-li.github.io/assets/docs/GenerativeAI

24、atWork.pdfMost of our respondents(51%)also believe that the skills gap and inequality could widen with AI,with the pessimism increasing with age.But even among the younger respondents the view that GenAI will sharpen disparities clearly prevails:respondents clearly fear that the smart would get smar

25、ter while the rest will be left behind(Figure 3).However,research from academics such as Erik Brynjolfsson,one of the leading researchers in the digital economy,suggests that GenAI could actually bridge the performance and learning gap between high-and low-skilled workers.Using data from 5,179 custo

26、mer support agents,Brynjolfsson and his team found that access to AI tools increases productivity(measured by issues resolved per hour)by 14%on average,including a 35%improvement for novice and low-skilled workers,albeit with minimal impact on experienced and highly skilled workers.Brynjolfsson and

27、his co-authors provide evidence that the AI model disseminates the best practices of more able workers and helps newer workers move down the experience curve.Additionally,they found that AI assistance improves customer sentiment,increases employee retention and may lead to worker learning.3 Therefor

28、e,the adoption of GenAI should have a levelling effect on wages,at least within individual companies.26%36%30%22%19%51%47%50%52%53%24%17%20%27%28%0%20%40%60%80%100%TotalGen ZMillennialGen XBoomersI dont knowAI will widen the gap between high-skilled experts and the rest of the workforce income inequ

29、ality will increaseAI will usher in a productivity boom,lifting wages across the board income inequality will decreaseFigure 3:Who will benefit from AI?In your opinion,how will AI affect the wages in your country?Will AI increase or decrease income inequality?Share of respondents,in%Source:Allianz R

30、esearchAllianz Research8As a data-driven industry like few others,insurance could benefit from a wide variety of AI applications and use cases in that could drive up productivity.For example,predictive analysis could support marketing strategies,with personalized outreach through new channels.Real-t

31、ime analysis and big-data analytics could enhance product development and enable usage-based insurance and risk-detection services.Automated processes and tailored product recommendations could enhance sales and distribution.Enhanced engagement would improve customer service and policy administratio

32、n,leading to expedited assessment and settlement of claims,as well as improved fraud detection,and in turn to fairer and more cost-effective insurance.All this would ultimately lead to greater customer satisfaction,increased efficiency and potentially new revenue streams(Table 1).AI and the insuranc

33、e industry16 July 20249 9Table 1:Use cases of artificial intelligence across the insurance value chain.MarketingProduct DevelopmentSales&DistributionUnderwritingCustomer service&policy administrationClaims management Predictive analytics Analytics of Tailored product Image analysis Predictive analys

34、is Prediction of claims Automated demand customer advice Natural language Voice recognition patterns analysis preferences Sales process processing(NLP)NLP Image recognition Product innovation automation Risk prevention and Anomaly/fraud mitigation detection New marketing Accurate pricing Reducing sa

35、les costs,Improved Personalized service Accurate claims channels Tailored products this enhancing quality/speed Improved customer assessments Tailored outreach Rapid product affordability of risk analysis engagement Fraud reduction adjustment including Increased resilience Faster responses complex r

36、isks of insuredsUse casesBenefitsSources:The Geneva Association,adapted from Eling et al.and Accenture,Allianz ResearchKey in reaping these gains is not to replace employees by AI tools but using AI to complement and augment their skills.This requires massive investments in re-skilling,keeping emplo

37、yees in the loop and preparing them for new kinds of work.First and foremost,AI adoption is not about cost reduction but creating new customer experiences.Increased productivity is a corollary.But this will be a huge challenge.So far,not many companies seem to have managed new technologies in that w

38、ay,as evidenced by the macro data:Despite the digital transformations of the last few decades,overall labor productivity has been flatlining in Western Europe since before the pandemic.To measure productivity in the insurance industry specifically,we take gross written premiums in euros for the tota

39、l market(life,health,and property and casualty)per number of employees in the insurance industry in our selected economies,using data from the CEDEFOP(European Centre for the Development of Vocational Training)and our annual Allianz Global Insurance Report.We find that labor productivity has risen i

40、n all countries when comparing 2010 to the end of 2023,but the increase has been more marked in Germany,Spain and Austria than in Italy,Poland and Spain(Figure 4).Looking ahead,using the labor forecasts from the CEDEFOP and our own insurance outlook,we expect insurance market productivity growth in

41、general to accelerate,with southern European countries to outpace the growth of their northern neighbors.This is in part the result of our forecast models which predict a recovery of life markets,driven by higher rates,slightly benefiting the southern European markets that were hit hard during the p

42、revious decade of ultra-low yields.Then,when assessing how the increased productivity prompted by GenAI technologies could affect the labor market(see Appendix 2 for methodology),we find that a 0.622%increase in productivity would yield a 1%decrease in labor to keep the insurance output as gross wri

43、tten premium stable.However,it is worth keeping in mind that the estimates of productivity growth from AI could be exaggerated,based on early evidence of easy-to-learn tasks.Some of the future effects will come from hard-to-learn tasks,where there are many context-dependent factors affecting decisio

44、n-making and no objective outcome measures from which to learn successful performance.Consequently,the economist Daron Acemoglu predicts that the total factor productivity gains over the next 10 years could be more modest at less than 0.55%.Allianz Global Insurance Report 2024:Transformative years a

45、head for the insurance sector.Allianz Research10Figure 4:Better times aheadGross written premiums in EUR,as a ratio of employees in insurance companiesSources:CEDEFOP,Allianz ResearchFinally,AI could also help the insurance industry,among others,face the looming challenge of a shrinking workforce am

46、id aging populations(Figure 5).Our previous research has shown that the working age population in the EU-27 will shrink by 20%until 2050.Italy,Spain and Germany will be hit even harder by demographic change.Banking on migration alone would require inflows of between 100,000 and 500,000 working migra

47、nts per year in the four largest economies alone.Against this backdrop,the acceleration and adoption of automating tools in the EU could free up some of the labor force to reskill towards industries that will need more workers,such as healthcare,STEM-related professions and other high-skill roles.-1

48、,000,000 2,000,000 3,000,000 4,000,000 5,000,000 6,000,000AustriaFranceGermanyItalyPolandSpain European labor markets:Migration matters16 July 20241111Figure 5a:The shrinking middleWorking-age population(15+)by age group,share of total in%Sources:CEDEFOP,Allianz Research0510152025303515-1920-2425-29

49、30-3435-3940-4445-4950-5455-5960-6415-1920-2425-2930-3435-3940-4445-4950-5455-5960-6415-1920-2425-2930-3435-3940-4445-4950-5455-5960-6415-1920-2425-2930-3435-3940-4445-4950-5455-5960-6415-1920-2425-2930-3435-3940-4445-4950-5455-5960-6415-1920-2425-2930-3435-3940-4445-4950-5455-5960-64AustriaFranceGe

50、rmanyItalyPolandSpain203520232013Figure 5b:The shrinking totalWorking-age population 25 to 64 years old,index 2010=100859095100105110AustriaFranceGermanyItalyPolandSpainSources:CEDEFOP,Allianz ResearchAllianz Research12It would be nave to assume that all new technological advances can only increase

51、welfare.In fact,Dan Ariely,a prominent behavioral scientist,purports that GenAI is like a highway:you can go faster,but you can also be stuck in a traffic jam.Some of the new tasks created by AI may indeed bring no competitive advantage if adoption is widespread or they may even have a negative soci

52、al value if used for deep fakes or internet manipulation.Collingridges Dilemma states that forecasting technological changes is a two-fold problem.On the one hand,there is an information challenge as impacts cannot be predicted until the technology is extensively developed and widely used.The second

53、 is a power issue as once the technology is entrenched in society controlling for it or developing policies to contain it is difficult.Although this might be true,we should at the very least prepare for the known unknowns.48%of the respondents in our survey believe that strict regulation is indispen

54、sable in the wake of the GenAI revolution,while 32%would prepare to safeguard Europes competitiveness in this important field.This is a delicate balance that policymakers as well as the private sector need to navigate,even more so as less than 10%see a laissez-faire approach as the best way forward.

55、GenAI is too important to leave its development to the tech giants alone(Figure 6).The way forward16 July 20241313Figure 6:Who calls the shots in AI?AI is a general-purpose technology;it can do harm or good.Therefore,governments should strictly regulate it to protect us from harmful applications.Sha

56、re of respondents,in%Source:Allianz Research48%47%47%45%46%46%55%32%32%31%33%34%31%29%9%8%9%11%7%10%7%12%14%13%11%12%13%10%0%50%100%TotalAustriaFranceGermanyItalyPolandSpainI dont knowNo,AI is a new technology,the market will find out how to use it best regulation would only curtail the development

57、of AI,limiting its potentialYes,regulation is necessary but only light regulation to safeguard Europes competitiveness in this important fieldYes,fully agree strict regulation is indispensable.Allianz Research14Appendix 1:Overall responsibility for methods:Allianz Research,Allianz SE Planning and dr

58、awing the sample:Qualtrics Target groups surveyed:Austrian resident population,age 18 and over in Austria French resident population,age 18 and over in France German resident population,age 18 and over in the Federal Republic of Germany Italian resident population,age 18 and over in Italy Polish res

59、ident population,age 18 and over in Poland Spanish resident population,age 18 and over in Spain Number of respondents:6,271 persons(1,172 from Austria,1,020 from France,1,020 from Germany,1,021 from Italy,1,032 from Poland and 1,006 from Spain)Sampling method:Representative quota sampling:Qualtrics

60、was given quotas for how many people to survey and which criteria to use in selecting respondents.The quotas were distributed in accordance with official statistics among sex,age groups and education.Representativeness:A comparison with official statistics shows that the survey data on the whole cor

61、responds to the total population age 18 and over in the three countries.Type of survey:Web-based survey Date of survey execution:26 April 2024 to 13 May 202416 July 2024Allianz Research15Appendix 2:To assess how an increase in productivity caused by the introduction of augmentation and complementary

62、 capabilities with GenAI technologies would affect the labor market,we used a simple a reduced-form fixed effects panel data model under the following specification:Laborit=0+1Productivityit+2Wageit+3UnemploymentRateit+EitWhere:Laborit is the number of employees in the insurance industry in country

63、i at time t.Productivityit is the gross written premium per employee in country i at time t.Wageit is the estimated average wages in the insurance industry in country i at time t.UnemploymentRateit is the unemployment rate in country i at time t.E is the error term.Although initially we added the wo

64、rking-age population growth variable to our model specification,it was not statistically significant and we dropped it.The key coefficient we focused on was the coefficient of our measure of labor productivity.Table 2:Panel data fixed effects regression coefficientsLOG(PRODUCTIVITY)-0.622SKILLS0.004

65、UNEMPLOYMENT-0.008LOG(WAGES)0.216FIXED EFFECTS,in%Source:Allianz Research16ALLIANZ RESEARCHteamOur16 July 2024Chief Economist Allianz SELudovic SAna Boataana.boataallianz-Arne HHead of Economic Research Allianz TradeHead of Insurance,Wealth&ESG ResearchAllianz SEFranoise HuangSenior Economist for As

66、ia Pacificfrancoise.huangallianz-Manfred StamerSenior Economist for Middle East&Emerging Europemanfred.stamerallianz-Luca MonetaSenior Economist for Emerging Marketsluca.monetaallianz-Macroeconomic ResearchMaxime LemerleLead Advisor,Insolvency Research maxime.lemerleallianz-Ano KuhanathanHead of Cor

67、porate Researchano.kuhanathanallianz-Corporate ResearchMichaela GrimmSenior Economist,Demography&Social PKathrin StoffelEconomist,Insurance&WPatricia Pelayo-RomeroSenior Economist,Insurance&ESGpatricia.pelayo-Insurance,Wealth and Trends ResearchPablo Espinosa UrielInvestment Strategist,Emerging Mark

68、ets&Alternative Assetspablo.espinosa-Capital Markets ResearchMarkus ZimmerSenior Economist,ESGJordi Basco CarreraLead Investment Strategistjordi.basco_Maria LatorreSector Advisor,B2Bmaria.latorreallianz-Maxime Darmet CucchiariniSenior Economist for US&Francemaxime.darmetallianz-Maddalena MartiniSeni

69、or Economist for Italy,Greece&BJasmin GrschlSenior Economist for EBjoern GriesbachSenior Investment Strategist&Eurozone EPatrick HoffmannEconomist,ESG&AIYao LuSector Advisoryao.luallianz-Lluis DalmauEconomist for Africa&Middle Eastlluis.dalmauallianz-Hazem KricheneSenior Economist,C18Recent Publicat

70、ionsDiscover all our publications on our websites:Allianz Research and Allianz Trade Economic Research12/07/2024|What to watch 11/07/2024|Olympic Games 05/07/2024|What to watch 03/07/2024|Securing critical infrastructure:the two-for-one of green investment28/06/2024|What to watch 25/06/2024|Mid-year

71、 Economic Outlook 2024-25:Games wide open?21/06/2024|What to watch 19/06/2024|Industrial policy:old dog,new tricks?14/06/2024|What to watch 13/06/2024|Climate change and the double impact of aging06/06/2024|What to watch 04/06/2024|What to expect from the European elections 31/05/2024|What to watch

72、29/05/2024|Allianz Pulse 2024:What unites and separates the demos of Europe24/05/2024|What to watch 23/05/2024|Allianz Global Insurance Report 2024 16/05/2024|What to watch 14/05/2024|Trade Survey 202403/05/2024|What to watch 30/04/2024|Ashes to ashes,carbon to soil26/04/2024|What to watch 22/04/202

73、4|Global outlook for private debt&private equity:private(r)for longer?18/04/2024|What to watch17/04/2024|Latin America:Shall we dance?11/04/2024|The best is yet to come11/04/2024|What to watch05/04/2024|What to watch26/03/2024|Economic Outlook:Its a wrap!22/03/2024|What to watch21/03/2024|Global aut

74、o outlook:Steering through turbulence14/03/2024|What to watch13/03/2024|Trumponomics:the sequel07/03/2024|What to watch 06/03/2024|When the penny drops-analyzing longevity literacy in six countries 29/02/2024|What to watch 28/02/2024|Global insolvency outlook:Reality check22/02/2024|What to watch 16

75、/02/2024|What to watch 16 July 20241919Director of PublicationsLudovic Subran,Chief EconomistAllianz ResearchPhone+49 89 3800 7859Allianz Group Economic Researchhttps:/ 28|80802 Munich|GallianzallianzAllianz Trade Economic Researchhttp:/www.allianz- Place des Saisons|92048 Paris-La-Dfense Cedex|Fran

76、ceresearchallianz-allianz-tradeallianz-tradeAbout Allianz ResearchAllianz Research encompasses Allianz Group Economic Research and the Economic Research department of Allianz Trade.Forward looking statementsThe statements contained herein may include prospects,statements of future expectations and o

77、ther forward-looking statements that are based on managements current views and assumptions and involve known and unknown risks and uncertainties.Actual results,performance or events may differ materially from those expressed or implied in such forward-looking statements.Such deviations may arise du

78、e to,without limitation,(i)changes of the general economic conditions and competitive situation,particularly in the Allianz Groups core business and core markets,(ii)performance of financial markets(particularly market volatility,liquidity and credit events),(iii)frequency and severity of insured lo

79、ss events,including from natural catastrophes,and the development of loss expenses,(iv)mortality and morbidity levels and trends,(v)persistency levels,(vi)particularly in the banking business,the extent of credit defaults,(vii)interest rate levels,(viii)currency exchange rates including the EUR/USD

80、exchange rate,(ix)changes in laws and regulations,including tax regulations,(x)the impact of acquisitions,including related integration issues,and reorganization measures,and(xi)general competitive factors,in each case on a local,regional,national and/or global basis.Many of these factors may be mor

81、e likely to occur,or more pronounced,as a result of terrorist activities and their consequences.No duty to updateThe company assumes no obligation to update any information or forward-looking statement contained herein,save for any information required to be disclosed by law.Allianz Trade is the trademark used to designate a range of services provided by Euler Hermes.