《印孚瑟斯(Infosys):2023北美生成式AI雷达报告(英文版)(54页).pdf》由会员分享,可在线阅读,更多相关《印孚瑟斯(Infosys):2023北美生成式AI雷达报告(英文版)(54页).pdf(54页珍藏版)》请在三个皮匠报告上搜索。
1、GENERATIVE AI RADAR 2023 NORTH AMERICA Generative AI Radar 2023:North America|3External Document 2023 Infosys Limited Knowledge InstituteContentsExecutive summary 5Section 1 Optimism and high expectations 7Generative AI investment escalates 7The nimble giant 8Section 2 Generative AI in action 11Visi
2、ble impact on outcomes 11Rethinking experiences 12Healthcare,life sciences,and financial services lead 13Section 3 Leadership and talent 15Leaders care,and it shows 15The three paths to AI talent 17Section 4 Overcoming adoption challenges 19Data quality 20Talent and business model 20Ethics 22Conclus
3、ion 24Appendix Generative AI by industry 26Appendix Research approach 34Executive summaryKnowledge Institute4|Generative AI Radar 2023:North America External Document 2023 Infosys Limited Generative AI Radar 2023:North America|5External Document 2023 Infosys Limited Knowledge InstituteGenerative AI
4、burst on to the scene only recently,yet enterprises are already aggressively exploring its transformative potential.Unlike previous innovations such as blockchain or metaverse,consumer versions of generative AI are widely available and accessible,led by OpenAIs ChatGPT and DALL-E,as well as Midjourn
5、ey and Microsofts Copilot.The combination of lucrative benefits and early access has driven rapid enterprise adoption,reminiscent of Googles launch1 in1998.However,a large gap exists between employees tinkering with consumer tools and business value at scale,through authorization,integration,and for
6、mal use.Our research surveyed 1,000 US and Canadian businesses to understand the extent of generative AI implementation and early indications on its ability to deliver value.Generative AIs increasing momentumWe found that organizations are serious about generative AI,and the level of investment refl
7、ects this.In the 12 months since ChatGPTs release,we estimate that companies in Canada and the US have invested$3.3 billion in generative AI initiatives,and we forecast this to jump 67%to$5.6 billion in 2024.This view is further reinforced by the prevalence of CEOs and boards of directors who sponso
8、r and govern generative AI initiatives.Unlike previous innovations,executives cite C-suite alignment and funding as the least of their AI deployment challenges.AI and the nimble giantAlso surprising,we found companies with more than$10 billion revenue are more likely to adopt generative AI and claim
9、 business value from these initiatives.This contradicts the traditional stereotype that smaller,more nimble organizations adopt new technology faster.Larger companies seem to outpace smaller ones to extract value due to low entry barriers and opportunity costs.Another surprise is that adoption level
10、s are highest in the healthcare and life sciences sectors,closely followed by financial services.These highly regulated industries tend to be laggards2 in digital and innovation adoption.This is particularly revealing when compared to high-tech companies,which rank fourth in adoption yet do show the
11、 highest likelihood to generating value from generative AI.From hype to realityHype is one area where generative AI certainly aligns with expectations and the potential for disillusionment if the technology fails to meet initial inflated expectations.Interestingly,most businesses see generative AI a
12、s a tool for business growth,efficiency,and improving user experience and personalization outcomes that have yet to be proven at scale.Conversely,few businesses see it as a tool for content creation and creativity its leading consumer application to date.Combine this expectation mismatch with data c
13、hallenges as well as ethics and bias risks,and we predict many businesses will face a reality check in the coming year.Knowledge Institute6|Generative AI Radar 2023:North America External Document 2023 Infosys Limited Generative AI Radar 2023:North America|7External Document 2023 Infosys Limited Kno
14、wledge InstituteSection 1 Optimism and high expectationsGenerative AI investment escalatesBusiness leaders in the US and Canada are optimistic that generative AI as part of the larger AI juggernaut has the potential to transform their companies.Many already spend significantly to integrate generativ
15、e AI into their enterprises.These investments will only escalate in 2024,the so-called“Year Two”of generative AI,according to the business leaders we interviewed.In generative AIs break-out year,just over a quarter of the companies in our survey(26%)budgeted at least$5 million on generative AI initi
16、atives(Figure 1).Yet a significant group merely dipped their toes into the water:nearly half(45%)spent less than$1 million on generative AI in the past year,while relatively few(5%)invested more than$10 million.Looking ahead,spending is certain to increase significantly and almost universally.A larg
17、e majority of respondents(72%)said they plan to spend more on generative AI in the next 12 months,with none planning to cut back.And in many cases,they are flipping the most recent spending ratio,with very few companies only dabbling in the technology.In the coming year,just 13%of executives said th
18、ey intend to spend less than$1 million on generative AI,while one-fifth said they will invest more than$10 million.Overall,we expect that companies in the US and Canada will invest$2.3 billion more on generative AI in 2024 than 2023.Infosys used its survey responses covering companies of various siz
19、es and industries and extrapolated their spending into the future.Conservatively,we estimate that companies in the US and Canada have invested$3.3 billion in generative AI in the past 12 months.Based on what executives told us,generative AI spending is expected to grow to nearly$5.6 billion a 67%inc
20、rease.8|Generative AI Radar 2023:North America External Document 2023 Infosys Limited Knowledge InstituteFigure 1.Generative AI investments set to accelerate in the coming yearMore than$10M:4.6%More than$10M:20.2%$5M to$10M:20.9%$5M to$10M:27.4%$1M to$5M:29.2%$1M to$5M:39.4%Less than$1M:27.0%Less th
21、an$1M:13.0%No spending:18.3%N=992.Spending on generative AI is expected to increase.Almost three-quarters of respondents spent less than$5 million on generative AI in the last 12 months.None of our respondents expected spending on generative AI to decrease in the next 12 months.Last 12 monthsNext 12
22、 monthsGenerative AI spendingSource:Infosys Knowledge InstituteThe nimble giant Generative AI wins the prize as the most hyped technology in the world today.But these dynamics differ from“disruptive”innovations of the past,such as blockchain or voice-activated tech,that lacked major uptake by large
23、companies.Often,nimble startups adopt new technologies and show tantalizing hints of the value to come.Meanwhile,large corporations wait and see how these innovations progress,or they simply are too large to turn on a dime.This time,the largest companies are ahead of their smaller counterparts(Figur
24、e 2).We found that 73%of companies with more than$10 billion in revenue have implemented generative AI solutions.That is true of barely more than a third(38%)of the companies that have between$1 billion and$10 billion in revenues.Even fewer of the smallest companies we spoke to have implemented gene
25、rative AI(27%).Generative AI Radar 2023:North America|9External Document 2023 Infosys Limited Knowledge InstituteFigure 2.Large companies more likely to report business value from their generative AI initiativesSource:Infosys Knowledge InstituteThese industry giants not only spend more but also gene
26、rate more return from their investments.Around 30%of the larger companies we spoke to said that they deliver business value from their deployments,compared to less than 10%of the smaller firms.These dynamics are the reverse of what we typically see from emerging technologies and for good reasons.Gen
27、erative AI first emerged as a consumer tool,and has been widely available and straightforward to deploy.Even companies with ingrained processes could quickly and relatively easily incorporate generative AI into their workflow,at least on pilot projects.As a result,this innovation became surprisingly
28、 less risky for the traditionally risk averse.Rather than replacing existing processes,generative AI can be seen as something thats added to existing tools and technologies.In other words,there is less opportunity cost for large firms(with capacity)to experiment with generative AI without impeding e
29、xisting work,where a leaner,smaller company(without spare capacity)halts one initiative to test generative AI.In this context,it is logical that the largest companies more effectively apply their economies of scale.More than$10B$1B to$10B$500M to$1B3.6%0.3%1.5%16.5%34.8%36.9%18.9%7.8%45.5%28.7%9.0%2
30、3.7%42.0%30.6%Does not have generative AI initiatives and does not plan to implementDoes not have generative AI initiatives currently but is planning to implementIs experimenting with generative AI and/or defning proofs of conceptHas implemented,or is currently implementing,generative AI solutionsHa
31、s established generative AI use cases that create business valueRespondents(%),N=1,000.Knowledge Institute10|Generative AI Radar 2023:North America External Document 2023 Infosys Limited Generative AI Radar 2023:North America|11External Document 2023 Infosys Limited Knowledge InstituteSection 2 Gene
32、rative AI in actionVisible impact on outcomesThe rapid adoption of generative AI and the billions of dollars invested indicate that business leaders expect it to have a massive impact,perhaps even becoming a transformative technology.Indeed,when asked whether generative AI will provide a positive or
33、 negative impact on business outcomes,88%of respondents expected positive impacts on revenue,with 84%expecting the same for profit,83%for cost efficiency,and 82%for business model improvement(Figure 3).Figure 3.Generative AI business impactSource:Infosys Knowledge InstituteTalent7.2%72.4%20.4%Busine
34、ss model6.1%82.2%11.7%Cost efciency5.6%83.2%11.2%Proft4.7%83.7%11.6%Revenue3.4%88.0%8.6%Reputation2.5%79.9%17.6%Negative impactPositive impactNeutral impactRespondents(%),N=1,000.Business leaders we spoke to identified a surprising list of use cases where they believe generative AI will generate imp
35、act.Generative AI is widely viewed as a content generation tool,though with varying degrees of demonstrated success.While some news outlets have embarrassed themselves by publishing AI-generated articles that contained glitchy3,inaccurate4,and offensive5 material,elsewhere,individuals have prompt-en
36、gineered images of gorgeous,realistic nature scenes6.12|Generative AI Radar 2023:North America External Document 2023 Infosys Limited Knowledge InstituteHowever,our survey found that business leaders generally do not prioritize these capabilities as the best use of generative AI(Figure 4).Just 13%sa
37、id that content and creativity were generative AIs main applications.Instead,42%expect user experience and personalization to have the greatest impact.Executives also said that generative AI will be used most often to improve operational efficiency and automation(26%)or streamline product developmen
38、t and design(20%).Figure 4.Where companies expect generative AI to have the most impactSource:Infosys Knowledge InstituteImproved content generation and creativity Respondents(%),N=1,000.Streamlined product development and design Increased operational efciency and automation Enhanced user experience
39、 and personalization 12.7%20.2%25.5%41.6%Business leaders might be cautious due to generative AIs inconsistent performance in creative tasks,especially in public-facing written content.It may be competent enough to pass graduate school exams7.But on other occasions,it might create fictionalcourt cas
40、es8 and precedents for a legal brief.This unpredictability injects enough risk to make executives and PR staff cringe.Also,it can be difficult and time consuming to fact check AI output.For creative applications,generative AI tends to offer speed at the cost of quality.Rethinking experiencesGenerati
41、ve AIs expected value in user experience suggests a change in how we define this technology.Rather than an isolated writing image,music or code-generating tool,it can be seen as a personalized AI assistant with a range of customized skills synced to employee needs.Infosys has long argued that the mo
42、st valuable use for AI in general is to augment humans instead of replacing them9.We see particular promise for generative AI in the software engineering cycle.For those focused on project planning and analysis,an AI assistant could help with effort estimation,risk assessment,and simulations.Softwar
43、e testers can use generative AI to optimize the number and value of tests,eliminate redundancies,automate test script generation,and promote the self-healing of scripts.User experience also applies to customers with generative AI chatbots or other tools.Amazon offers generative AI to its sellers10 t
44、o assist with the creation of product Generative AI Radar 2023:North America|13External Document 2023 Infosys Limited Knowledge Institutedescriptions,titles,and listing details.User experience also applies within companies,where generative AI chatbots manage first-line IT support queries,for example
45、.Generative AI creates actionable insights from maintenance logs and automates human workflows.This approach not only allows companies to reimagine user experiences,but also enhances operational efficiency.This is evident in the fact that a fifth of the companies in our survey anticipate the most po
46、sitive impact of generative AI to streamline work and automation(Figure 4).Healthcare,life sciences,and financial services lead Our research shows that most large businesses take generative AI seriously,but not all industries are equally advanced.The financial services,healthcare,and life sciences s
47、ectors all data and tech focused are particularly keen to use this new technology(Figure 10 see Appendix A).However,they are traditionally cautious with new innovations due to their high level of regulatory scrutiny.In healthcare,generative AI can be used to create synthetic data11 for clinical rese
48、arch and medical education.Synthetic data could help avoid patient privacy concerns,a significant benefit for healthcare firms although see our discussion later about some of the concerns with synthetic data.Firms such as Paige12 also use generative AI to accelerate cancer diagnosis and reduce detec
49、tion errors.The firm is now collaborating with Microsoft to create the worlds largest AI models for digital pathology and oncology.Financial services firms also use generative AI to create synthetic data for training purposes13 although there are challenges with the use of synthetic data,which we di
50、scuss in more detail later in this report.In this case,the output data simulates fraudulent and regular financial transactions.Machine learning models then train on both the real and synthetic data to improve their real-world accuracy.Using real data and documents,investment bank Goldman Sachs14 has
51、 experimented with generative AI to classify and categorize its content library.Its worth noting that while these three sectors have implemented generative AI at a higher rate than others,the high-tech sector leads the way in the number of generative AI implementations delivering business value.The
52、financial services sector does come a close second but healthcare and life sciences lag far behind in this regard.Knowledge Institute14|Generative AI Radar 2023:North America External Document 2023 Infosys Limited Generative AI Radar 2023:North America|15External Document 2023 Infosys Limited Knowle
53、dge InstituteSection 3 Leadership and talentLeaders care,and it showsOne clear positive in this rapid embrace of generative AI is that it is being taken seriously by senior executives,and beyond the confines of the IT teams.Typically,the chief information officer(CIO)is the primary sponsor for techn
54、ology innovation programs,followed by business function leaders.Our data reflects this trend,with CIOs sponsoring generative AI initiatives in 26%of businesses surveyed.However,our survey also reveals that both chief executive officers(CEOs)and chief information security officers(CISOs)play leading
55、roles.CISOs are lead sponsors of these initiatives in 18%of businesses,and CEOs in 16%of the businesses we spoke to.As discussed later in this report,CEO involvement is a strong indicator that generative AI projects will deliver transformative outcomes not least Figure 5.Who is the main sponsor of g
56、enerative AI?Source:Infosys Knowledge InstituteChief information ofcerCybersecurity/chief information security ofcer25.6%Chief marketing ofcerRespondents(%),N=1,000.Chief executive ofcerBusiness unit leaderTo be determinedBoard of directorsChief operating ofcer0.3%16.0%13.2%10.1%9.6%7.3%17.9%16|Gene
57、rative AI Radar 2023:North America External Document 2023 Infosys Limited Knowledge Institutebecause the top-level leaders set strategic priorities,commit sufficient resources,and measure executives on program success.The significant CISO involvement finding is also positive,as it shows that many bu
58、sinesses take seriously the potential threats posed by generative AI.Security and privacy concerns extend beyond internal processes and policies to a rapidly evolving cyber-threat and regulatory landscape.Indeed,our survey also found significant scrutiny of generative AI regulations and policies ano
59、ther positive sign for the responsible rollout of this new and transformative technology.The CIO most often takes the lead on these initiatives(22%).However,the board of directors and CISO(both 21%)are just as likely to define generative AI regulations and policies for their companies(Figure 6).This
60、 reflects how seriously organizations view generative AI and the risk level they perceive,whether reputation,investment size,or transformation potential.Figure 6.Tech leaders and boards of directors most often lead generative AI governanceSource:Infosys Knowledge InstituteRespondents(%),N=1,000.Chie
61、f information ofcer or otherwise within ITBoard of directorsCybersecurity/chief information security ofcerTo be determinedWithin business units and individual product ownersChief operating ofcerLegal21.6%20.8%20.5%12.9%12.6%6.4%5.2%Generative AI Radar 2023:North America|17External Document 2023 Info
62、sys Limited Knowledge InstituteFigure 7.How companies plan to close the generative AI skills gapSource:Infosys Knowledge InstituteThe three paths to AI talentAs with most technologies and initiatives,talent is central to realizing the potential of generative AI.Those who scale AI across the organiza
63、tion will be in high demand,as will new skillsets such as prompt engineering and model fine-tuning.“While some worry that AI will take their jobs,someone who is an expert in AI will certainly do so,”Jensen Huang,CEO of NVIDIA,recently told Infosys15.As we discuss in our Tech Navigator16 report,AI ta
64、lent is one of the top four challenges for executives to transform their enterprise to an AI-first operating model.Our survey backs this up,revealing that lack of skills and talent is the biggest challenge for 18%of companies(Figure 8 see next page).We also found that most companies plan to tackle t
65、he skills challenge by upskilling and reskilling employees(41%).The next most often cited plan by leaders is to partner with vendors to leverage their skills and talent(33%).A smaller number(26%)plan to recruit talent with generative AI skills(Figure 7).Whether upskilling,partnering,or recruiting,th
66、is AI talent must combine requisite engineering knowledge with the softer skills necessary to work in small,cross-functional,product-based teams emphasizing characteristics such as empathy,problem-solving,and integrity.Respondents(%),N=1,000.Upskilling and/or reskillingRecruiting talent with generat
67、ive AI skillsPartner with vendors to leverage their skills and talent40.9%33.3%25.8%Knowledge Institute18|Generative AI Radar 2023:North America External Document 2023 Infosys Limited Generative AI Radar 2023:North America|19External Document 2023 Infosys Limited Knowledge InstituteSection 4 Overcom
68、ing adoption challengesThe high-profile sponsorship of generative AI allows companies to avoid the ultimate barrier to implement new technologies:change management at the top.With new innovative technologies,top executives often either dont recognize the potential of new technologies or dont priorit
69、ize their adoption.Fortunately,generative AI does not have that problem.Only 4%of respondents cited C-suite buy-in among their most significant challenges.With executive support and growing budgets,generative AI has a clearer path to success than many other new technology initiatives.But that does n
70、ot guarantee a smooth road to adoption.Privacy concerns,data quality,and lack of talent all weigh on the minds of business leaders(Figure 8).In fact,these issues contribute to the“pilot purgatory”faced by companies struggling to move from small experimental projects to adoption at scale.Figure 8.Dat
71、a,skills and ethics are primary generative AI challengesSource:Infosys Knowledge InstituteRespondents(%),N=1,000.C-suite buy-inLack of investment Cannot access generative AI tools from corporate computers Concerns about ethics,bias,fairness,safety Lack of skills,knowledge,or resourcesData not in a u
72、sable stateData privacy and security25.9%23.4%18.4%14.4%8.1%5.4%4.4%20|Generative AI Radar 2023:North America External Document 2023 Infosys Limited Knowledge InstituteData qualityAlmost half of all respondents cite data challenges either privacy and security,or usability as their biggest obstacles
73、to generative AI implementation.“Hallucinations”and intellectual property(IP)infringement are significant inherent risks in using generative AI that relies on public data.For that reason,many businesses look to build organization-specific tools trained on corporate rather than public data.Yet corpor
74、ate data is often not full,complete,and formatted for effective use.Further,data quality is a particularly difficult problem to solve.One option is to use synthetic data,a process that uses statistical algorithms to fill the gaps in data sets.This approach can be used to model real-world situations
75、and train generative AI models and algorithms.However,it comes with several concerns,including the cost of building and keeping consistent with the original data.Synthetic data also tends to mimic and replicate the biases inherent in the original data that underlies the synthetic dataset.While in th
76、eory synthetic data does not contain personal information,it nonetheless is linked to real data about real people.Businesses considering synthetic data to build their own generative AI tools must work closely with data and ethics practitioners to manage these risks.All data,whether synthetic or real
77、,needs to achieve an acceptable state before it builds generative AI models and tools.Firms also need access to data scientists to clean and classify data before using in generative AI.Synthetic data and subsequently,models fed with generative AI-created information,risk creating a downward spiral o
78、f quality that will undermine the utility of and trust in public foundation models.Talent and business modelThe success of generative AI is challenged as much by talent as data science.As we discuss in Tech Navigator17,the Horizon Technology Innovation model is useful to map the stages of an organiz
79、ations AI journey,expressed as three horizons:H1,H2,and H3(Figure 9).The first wave is driven by machine learning,where data scientists with math and econometric skills are in demand.In H2,driven by deep learning,the need shifts to data engineers.Generative AI is an H3 technology,and a key talent re
80、quirement here is prompt engineers,who straddle the boundary between programming and creative writing.They are very much in demand and highly compensated.As we stated earlier in this report,many companies seek to upskill and reskill their current employees,though significant numbers also look outsid
81、e for new skills.Our research finds that while most Generative AI Radar 2023:North America|21External Document 2023 Infosys Limited Knowledge InstituteFigure 9.Three horizons companies cross in their AI journeySource:Infosys Knowledge Instituterespondents believe generative AI will have a positive i
82、mpact in all business outcomes,a small group expects to see negative effects on their talent,business models,or cost efficiency(Figure 3).As with many digital innovations,companies need to consider their holistic operating model.Based on our research,internal experience,and client work,an operating
83、model has emerged that harnesses AIs potential and addresses risk while evolving operations.While generative AIs initial outlook may be positive,success will require careful planning and design as talent and operating model constraints mount.H2Transfer learning,responsible AI(less data,explainable s
84、ystems)H1Conventional AI(augmenting intelligence)KEY PATTERNS AI governance AI ethics,explainable AI Model pruning,quantization tech Transfer learning Neural networks Object detection,classifcation,segmentation Prediction recommendations Logistic regression Classifcation regression Rule-based Expres
85、sion-basedH3Transformer architectures,foundation models,generative AI(self-supervised)AI models should be capable of learning and evolving on their own with minimal human intervention.These are rapidly gaining prominence among enterprises.Businesses are investing in AI systems that are capable of ma
86、king fast,transparent,and unbiased decisions.These systems are already mainstream,providing AI-powered assistance to business decisions.Billion/trillion parameter models Zero-shot learning Multitask learning Multimodal and multilingual Closed and open access models Responsible by design Auto ML22|Ge
87、nerative AI Radar 2023:North America External Document 2023 Infosys Limited Knowledge InstituteEthicsWhile our respondents voiced data and talent as their primary concerns,leaders should not overlook ethics,which was only mentioned by 14%as a top concern.AI ethics,bias,and model transparency are wid
88、ely discussed beyond the corporate worlds:consumers and shareholders are aware of these issues,even if they dont understand the underlying tech.They are concerned that opaque generative AI tools make decisions based on data with embedded biases,perpetuating societal disadvantages such as discriminat
89、ion against certain ethnicities or genders.Other enterprise challenges for companies using generative AI include malicious use of AI-generated malware and misinformation,as well as copyright issues.US courts have ruled18 that generative AI outputs cannot be copyrighted,while lawsuits are pending in
90、the US19 about the use of original works scraped to train foundation models.With copyright concerns,businesses need active governance to oversee employee use of consumer tools that encourage them to upload corporate IP and commercially sensitive material.Building and deploying governance and generat
91、ive AI oversight in the workplace requires a responsible-by-design approach,not a reactive,ad-hoc exercise,and with senior executive oversight.Knowledge InstituteGenerative AI Radar 2023:North America|23External Document 2023 Infosys Limited 24|Generative AI Radar 2023:North America External Documen
92、t 2023 Infosys Limited Knowledge InstituteOur research shows that businesses in the US and Canada are excited about generative AI committing significant investment because they see it as a driver of growth.It is laudable to see the direct involvement by senior executives,and that larger businesses a
93、nd important sectors such as healthcare,life sciences and financial services are leading theway.However,as with any technology at the early stage of the hype curve,there is risk of unfocused investment and skipped foundational steps.We expect that many lessons will be learned,quickly,about the true
94、value and application of generative AI in enterprise landscapes in the next year.There is no denying that significant challenges exist around data governance and usability,and currently there are no quick solutions.But leaders who correct those issues responsibly now will reap greater rewards in the
95、 long term.The rush to train,contract and hire AI skills will also disrupt the talent landscape,squeezing companies ability to advance with this technology in the short term.The pace of change must remain quick,without taking shortcuts to overcome the short-term obstacles.Nonetheless,this is an exci
96、ting time,potentially a generational sea-change event.US and Canadian firms are clearly keen to explore generative AI technologies and determine how to create value from them.Generative AI is surprising us in many ways not least in its adoption patterns.As with all new technologies,we expect the ini
97、tial flush of excitement to subside,and then renewed focused optimism,overcoming temporary disillusionment along the way.ConclusionKnowledge InstituteGenerative AI Radar 2023:North America|25External Document 2023 Infosys Limited 26|Generative AI Radar 2023:North America External Document 2023 Infos
98、ys Limited Knowledge InstituteAppendix A Generative AI by industryThis survey covers 12 sectors in the US and Canada,allowing us to compare how extensively and effectively each industry is using generative AI.Although most firms are investing in this technology,there is a wide range of investment be
99、tween sectors.While differences exist between industries in the frequency of use cases that create value,only the retail(and hospitality)and the logistics(and supply chain)sectors are statistically different from the others(Figure 10).These two sectors have significantly fewer respondents,which sugg
100、ests these industry results are viewed more directionally.Figure 10.Generative AI adoption and business value by sectorSource:Infosys Knowledge InstituteRespondents(%),N=1,000.TelecommunicationsRetail or hospitalityManufacturingLogistics or supply chainLife sciencesInsuranceHigh techHealthcareFinanc
101、ial servicesEnergy,mining,or utilitiesConsumer packaged goodsAutomotive29.6%28.0%22.5%15.1%21.1%31.6%52.4%23.5%21.1%15.5%29.2%24.6%39.4%10.6%41.9%14.0%17.7%1.6%21.3%18.5%20.9%2.2%30.6%16.1%Has implemented,or is currently implementing,generative AI solutionsHas established generative AI use cases tha
102、t create business valueGenerative AI Radar 2023:North America|27External Document 2023 Infosys Limited Knowledge Institute28|Generative AI Radar 2023:North America External Document 2023 Infosys Limited Knowledge InstituteAutomotivePotential use cases Autonomous vehicle training Generate virtual env
103、ironments and synthetic data for realistic simulations User personalization Power in-car personal assistants and predict preferred routes and dashboard settings Marketing optimization Develop customer-centric content and track marketing investments Product development Accelerate design and delivery
104、stages,as well as data synthesis and pattern detection Maintenance Generative AI can create predictive maintenance models built on data from myriad auto and truck parts,and can be used to generate logs and advisory notices for drivers Road freight Generative AI can be used to train autonomous trucks
105、,and to help plan the most effective and efficient delivery routes for goods34%Enhance user experiences and personalization Automotive:Expectations of most positive impact from generative AIAlmost four in 10(39%)of automotive respondents expect streamlined product development and design as highest p
106、ositive impact for generative AI.This is almost twice the proportion of the overall survey(20%).At the same time,significantly fewer automotive respondents(15%compared to 26%overall)felt that increases in operational efficiency and automation will be the primary generative AI value areas.15%Increase
107、 operational efficiency and automation39%Streamline product development and design11%Improve content generation and creativityGenerative AI Radar 2023:North America|29External Document 2023 Infosys Limited Knowledge Institute30|Generative AI Radar 2023:North America External Document 2023 Infosys Li
108、mited Knowledge InstituteOverall,the largest group of respondents expect generative AI to have the biggest positive impact in enhancing user experiences and personalization.This sentiment was even stronger among respondents in the consumer packaged Potential use cases Product descriptions Create SEO
109、-friendly and engaging product descriptions,enabling mass personalization QR codes as digital art Transform QR codes into visually appealing brand identifiers Inventory management Create logs and recommendations for stock management including demand forecasting,allocation of stock and risk assessmen
110、ts Display design Text-to-image tools can be used to generate marketing campaigns and packaging Visual merchandising Gather data from a store to track customer footfall and throughput to optimize store planograms CPG:Expectations of most positive impact from generative AIConsumer packaged goodsgoods
111、 industry,with more than half(54%)feeling this way compared to 42%overall.Conversely,far fewer respondents from this industry(9%)felt generative AI would have the most positive impact on operational efficiency and automation compared to more than one quarter(26%)across industries.54%Enhance user exp
112、eriences and personalization 9%Increase operational efficiency and automation22%Streamline product development and design16%Improve content generation and creativityGenerative AI Radar 2023:North America|31External Document 2023 Infosys Limited Knowledge Institute32|Generative AI Radar 2023:North Am
113、erica External Document 2023 Infosys Limited Knowledge InstituteOne in three respondents from the energy,mining,and utilities sectors expect enhanced user experiences and personalization to be the most important impact of generative AI.One in four respondents expects improved operational efficiency
114、and automation to be the best.This sentiment is in line with Potential use cases Grid management Analyze data such as consumption patterns and load distribution to optimize grid performance and predict problems Renewable energy Help in the integration of renewable sources,including solar and wind,by
115、 predicting output and optimizing resource allocation Demand forecasting Enhance the accuracy of demand forecasts,allowing for efficient resource allocation and cost savings Resource optimization Simulate mining scenarios to optimize resource allocation,potentially saving costs and time Ore quality
116、prediction Create models to predict the ore quality based on geological data Environmental impact modeling Simulate the environmental impact of various mining methodsour survey respondents overall.However,these sectors differ in the proportion of respondents that think the biggest impact of generati
117、ve AI will come from streamlined product development and design.More than a quarter(29%)of respondents felt this is where generative AI would shine,compared to just 20%across all sectors.Energy,mining and utilities:Expectations of most positive impact from generative AIEnergy,mining,and utilities33%
118、Enhance user experiences and personalization 25%Increase operational efficiency and automation29%Streamline product development and design12%Improve content generation and creativityGenerative AI Radar 2023:North America|33External Document 2023 Infosys Limited Knowledge Institute34|Generative AI Ra
119、dar 2023:North America External Document 2023 Infosys Limited Knowledge InstituteHealthcare and life sciences respondents were much more likely to think generative AI would have its greatest positive impact on operational efficiency and automation.This was over 50%of life sciences and healthcare res
120、pondents compared to a survey average of just 26%.On the other hand,these respondents were much less likely to think that the biggest positive impact from Potential use cases Drug,gene,and protein sequence design Accelerate drug discovery by designing molecules and proteins and optimize synthetic ge
121、ne design Personalized treatment plans Analyze a patients medical history to generate customized treatment plans Enhanced medical imaging Use AI algorithms to improve the accuracy of medical imaging techniques like computerized tomography and magnetic resonance imaging scans by automatically identif
122、ying abnormalities Patient triage Chatbots trained on specialized LLMs can provide first-line patient triage,assessing symptoms to send the patient to an appropriate human professionalgenerative AI would come from streamlined product development and design or improved content generation.Only about 1
123、 in 20 respondents here felt these were the areas that generative AI would do best.This is compared to 20%of all respondents for streamlining product development and 13%for content generation.Healthcare and life sciences:Expectations of most positive impact from generative AIHealthcare and life scie
124、nces35%Enhance user experiences and personalization 54%Increase operational efficiency and automation5%Streamline product development and design6%Improve content generation and creativityGenerative AI Radar 2023:North America|35External Document 2023 Infosys Limited Knowledge Institute36|Generative
125、AI Radar 2023:North America External Document 2023 Infosys Limited Knowledge InstituteHigh-tech respondents differed little in their opinions on user experience and product development.However,they were the most likely to think that generative AIs biggest positive impact would be increased efficienc
126、y Potential use cases Software creation Tools such as Microsofts Copilot are being used to help developers generate new code,review existing code and help with code completion Data analytics Generative AI can be used to generate reports from data analytics to highlight narrative points for storytell
127、ing and flag up anomalies and problems Software analysis Analyze completed code to identify bugs and suggest fixes,and analyze code for adherence to guidelines,ensuring consistency Business analysis Generative AI tools can eview code to create reports that can be used across the wider organization,i
128、ncluding business analysts and product managers Automation Generative AI tools can be used to automate repetitive tasks,freeing up humans to do more creative or higher-level workHigh tech:Expectations of most positive impact from generative AIHigh techand automation(34%).The overall average was 26%.
129、They were also the least likely to rank improved content generation and creativity as offering the biggest positive impact(5%),compared to 13%overall.43%Enhance user experiences and personalization 34%Increase operational efficiency and automation18%Streamline product development and design5%Improve
130、 content generation and creativityGenerative AI Radar 2023:North America|37External Document 2023 Infosys Limited Knowledge Institute38|Generative AI Radar 2023:North America External Document 2023 Infosys Limited Knowledge Institute Insurance and financial servicesRespondents from the insurance and
131、 financial services industries are generally in line with the survey average for where generative AI will have its greatest positive impact Potential use cases Document management LLMs trained on the firms corpus can provide document summaries Fraud detection Generative AI can dentify anomalies and
132、generate reports to flag up potential fraud Customer service Chatbots built on LLMs and trained on domain-specific data can manage interactions with customers from loan decisions to helping them report issues Training Create firm-specific and role-specific training materials,from videos and interact
133、ive scenarios to role-play exercises Employee compliance Create listening tools that can monitor email,voice telephony and messaging across a firm to spot any issues with compliance and reporting Content Create personalized marketing and website contentexcept in one area.Almost one quarter(23%)of in
134、surance and financial services respondents expect that generative AI will be most impactful in improving content generation and creativity.This is significantly higher than the survey average of 13%.Insurance and financial services:Expectations of most positive impact36%Enhance user experiences and
135、personalization 21%Increase operational efficiency and automation20%Streamline product development and design23%Improve content generation and creativityGenerative AI Radar 2023:North America|39External Document 2023 Infosys Limited Knowledge Institute40|Generative AI Radar 2023:North America Extern
136、al Document 2023 Infosys Limited Knowledge InstituteLogistics and supply chain managementSentiments in logistics and supply chain management were similar to those of consumer packaged goods.Many more respondents in this industry(66%)expect generative AI to excel at enhancing Potential use cases Rout
137、e optimization Assess and suggest the most efficient routes for delivery runs based on road conditions,weather,and driver availability Risk management Analyze data on geopolitical concerns,weather events,industrial unrest,and more to produce dashboards where workers can see issues at a glance,and ge
138、nerate reports to suggest mitigations Supplier management Facilitate interactions with suppliers and pull insights from proprietary data and external soruces such as news articles to refine understanding of trends that impact suppliers Price intelligence Gather information on competitor pricing and
139、costsuser experiences and personalization,compared to 42%overall.Further,just 3%of logistics and supply chain respondents expect generative AI to make product development and design more streamlined,a fraction of the 20%across sectors.Logistics and supply chain management:Expectations of most positi
140、ve impact from generative AI66%Enhance user experiences and personalization 16%Increase operational efficiency and automation3%Streamline product development and design15%Improve content generation and creativityGenerative AI Radar 2023:North America|41External Document 2023 Infosys Limited Knowledg
141、e Institute42|Generative AI Radar 2023:North America External Document 2023 Infosys Limited Knowledge InstituteManufacturingSentiments in manufacturing were similar to those of the survey as a whole.About the same proportion expected generative AI to shine in each of the four areas we asked about wh
142、en looking at responses across industries.Potential use cases Product design Accelerate the design of both entire products and of components Maintenance Monitor machinery failures and predict potential breakdowns,enabling proactive management of equipment Quality management Training an AI on images
143、of similar products and identifying those that were defective can help predict potential problems with a new product Inventory management Simulate scenarios such as weather-driven surges in demand and use historical data to fine-tine production schedules and optimize stock levels Cost management Usi
144、ng generative AI tools to manage design,maintenance,and risk can reduce production costs User feedback Large language models can pull together reports of feedback from customers to further optimize design Manufacturing:Expectations of most positive impact from generative AI42%Enhance user experience
145、s and personalization 21%Increase operational efficiency and automation28%Streamline product development and design9%Improve content generation and creativityGenerative AI Radar 2023:North America|43External Document 2023 Infosys Limited Knowledge Institute44|Generative AI Radar 2023:North America E
146、xternal Document 2023 Infosys Limited Knowledge InstituteRetail and hospitalitySentiments in retail and hospitality were similar to the survey overall.About the same proportion expected generative AI to excel in each of the four areas of potential benefit.Potential use cases Marketing Large language
147、 models can draft highly personalized content for customers to boost engagement and conversion Insights Generative AI tools can process data and generate reports for marketers,helping them build more effective campaigns Supporting content Creative teams can use generative AI tools to streamline maki
148、ng promotional videos,writing blogs,and creating other content to support retail campaigns Customer service/concierge services Chatbots built on LLMs and trained on domain-specific data can assist with orders,as well as with booking tables and tickets,and assistance for visitors with local informati
149、on Translation Generative AI can provide language translation services for visitors from overseas Energy management Analyse energy use patterns to manage energy use on retail premisesRetail and hospitality:Expectations of most positive impact from generative AI43%Enhance user experiences and persona
150、lization 20%Increase operational efficiency and automation28%Streamline product development and design9%Improve content generation and creativityGenerative AI Radar 2023:North America|45External Document 2023 Infosys Limited Knowledge Institute46|Generative AI Radar 2023:North America External Docum
151、ent 2023 Infosys Limited Knowledge InstituteTelecommunicationsTelecommunications respondents provided similar answers to the overall average,except in one area.Significantly fewer(11%)expected the most positive impact from generative AI to come from operational efficiency and automation.This is abou
152、t half the proportion of the survey average of 26%.The telecom sector is already rolling out generative AI for customer interactions.In Australia,Googles Dialogflow is being used by Optus,one of the largest telcos in the country,to power virtual agents for customer support.This technology comes with
153、 prebuilt agents,which can be rolled out quickly to deal with queries about bill payments and orders without requiring custom programming.Potential use cases Customer support Chatbots can help users with billing queries and orders Engineer support Generative AI can be trained on the network topology
154、 and guide engineers through tasks by providing interactive guidance Software development Help developers create specialized code and applications more quickly and efficiently Network optimization Analyze network data and conditions and generate insights to streamline resource deployment Network sec
155、urity Track threats and assess vulnerabilities by analyzing network traffic to identify malicious activityTelecoms:Expectations of most positive impact from generative AI52%Enhance user experiences and personalization 11%Increase operational efficiency and automation16%Streamline product development
156、 and design21%Improve content generation and creativityKnowledge InstituteGenerative AI Radar 2023:North America|47External Document 2023 Infosys Limited 48|Generative AI Radar 2023:North America External Document 2023 Infosys Limited Knowledge InstituteAppendix B Research approachInfosys commission
157、ed a survey of 1,000 companies in the US and Canada during August and September 2023 via telephone interviews to gauge their attitudes to and adoption of generative AI.The survey looked at 12 industries:automotive;consumer packaged goods,energy,mining,or utilities;financial services;healthcare;high
158、tech;insurance;life sciences;logistics or supply chain;manufacturing;retail or hospitality;and telecommunications.We asked respondents about the state of generative AI in their organizations,including questions about investment plans,how generative AI is rolled out and managed across the organizatio
159、n,where the leadership comes from in the business,and how confident the respondents are about the readiness of their company to adopt and use generative AI.We also asked about where respondents expect generative AI to have the most impact in their business,as well as questions about the company loca
160、tion and size.To determine the generative AI investment figures:We asked the range of spending on generative AI for each respondent.We used the midpoint(or the lower bound in the case of an infinite range)as an estimate for the amount spent on generative AI for each respondent.We then grouped respon
161、dents into our 12 industries and created totals for each industry.We then adjusted our industry spending totals based on the difference in industry representation in our sample,compared to the distribution of companies in reality,using data from Refinitiv.Finally,we calculated industry totals for bo
162、th trailing 12 months and the next 12 months.Generative AI Radar 2023:North America|49External Document 2023 Infosys Limited Knowledge InstituteTelecommunicationsRetail or hospitalityManufacturingLogistics or supply chainLife sciencesInsuranceHigh techHealthcareFinancial servicesEnergy,mining,or uti
163、litiesConsumer packaged goodsAutomotive6.2%9.1%10.8%6.2%8.6%6.6%6.5%8.4%9.8%11.4%9.3%7.1%Primary industryRespondents(%),N=1,000.Company headquartersUS80%Canada20%Respondents(%),N=1,000.Current job levelC-suite41.3%Executive(vice-president,senior vice-president,director)58.7%Respondents(%),N=1,000.Mo
164、re than$10B33.3%$1B to$10B33.4%$500M to$1M33.3%Global revenue(past 12 months)Respondents(%),N=1,000.50|Generative AI Radar 2023:North America External Document 2023 Infosys Limited Knowledge InstituteReferences1.From the garage to the Googleplex,n.d.,About Google.2.The technology imperative for life
165、 sciences,Mike Joyce,Jeffrey Lewis,Manuel Mller,and Grard Richter,January 30,2020,McKinsey&Company.3.USA Today owner pauses AI articles after butchering sports coverage,Maggie Harrison,August 20,2023,Futurism.4.CNET defends use of AI blogger after embarrassing 163-word correction:“Humans make mistak
166、es,too,”Maxwell Strachan,January 17,2023,Vice.5.Microsoft calls deceased NBA player“useless”in AI-written obituary,Nicole Agius,September 15,2023,Search Engine Land.6.What does generative AI mean for bird and nature photography?Allen Murabayashi,Summer 2023,Audubon.7.ChatGPT passes exams from law an
167、d business schools,Samantha Murphy-Kelly,January 26,2023,CNN Business.8.Lawyer cites fake cases generated by ChatGPT in legal brief,Lyle Moran,May 30,2023,Legal Dive.9.Tech navigator:The AI-first organization,Rajeshwari Ganesan,Rajeev Nayar,Kamalkumar Rathinasamy,Rafee Tarafdar,Kate Bevan,and Harry
168、Keir Hughes,2023,Infosys Knowledge Institute.10.Amazon launches generative AI to help sellers write product descriptions,Mary Beth Westmoreland,September 13,2023,About Amazon.11.Generative adversarial networks and synthetic patient data:Current challenges and future perspectives,Anmol Arora and Anan
169、ya Arora,July 2022,Future Healthcare Journal,Vol.9,pp.190-193.12.Home page,n.d.,Paige.Generative AI Radar 2023:North America|51External Document 2023 Infosys Limited Knowledge Institute13.Use cases of generative AI across different industries,September 15,2023,emizentech.14.Goldman Sachs CIO tests g
170、enerative AI,Isabelle Bousquette,May 2,2023,The Wall Street Journal.15.Tech navigator,2023.16.Tech navigator,2023.17.Tech navigator,2023.18.AI-generated art cannot receive copyrights,US court says,Blake Brittain,August 21,2023,Reuters.19.US judge finds flaws in artists lawsuit against AI companies,B
171、lake Brittain,July 20,2023,Reuters.52|Generative AI Radar 2023:North America External Document 2023 Infosys Limited Knowledge InstituteAuthors Samad Masood|Infosys Knowledge Institute,LondonKate Bevan|Infosys Knowledge Institute,LondonJeff Mosier|Infosys Knowledge Institute,DallasAnalysis and Produc
172、tionIsaac LaBauve|Infosys Knowledge Institute,DallasPramath Kant|Infosys Knowledge Institute,BengaluruPragya Raj|Infosys Knowledge Institute,BengaluruGenerative AI Radar 2023:North America|53External Document 2023 Infosys Limited Knowledge InstituteAbout Infosys Knowledge InstituteThe Infosys Knowle
173、dge Institute helps industry leaders develop a deeper understanding of business and technology trends through compelling thought leadership.Our researchers and subject matter experts provide a fact base that aids decision making on critical business and technology issues.To view our research,visit I
174、nfosys Knowledge Institute at or email us at .54|Generative AI Radar 2023:North America External Document 2023 Infosys Limited Knowledge Institute 2023 Infosys Limited,Bengaluru,India.All Rights Reserved.Infosys believes the information in this document is accurate as of its publication date;such in
175、formation is subject to change without notice.Infosys acknowledges the proprietary rights of other companies to the trademarks,product names and such other intellectual property rights mentioned in this document.Except as expressly permitted,neither this documentation nor any part of it may be repro
176、duced,stored in a retrieval system,or transmitted in any form or by any means,electronic,mechanical,printing,photocopying,recording or otherwise,without the prior permission of Infosys Limited and/or any named intellectual property rights holders under this document.For more information,contact I|NYSE:INFYStay Connected