人工智能siri搞笑对话

1000100:00:00,000-->00:00:03,177{\an8}人工智能下一次工业革命的催化剂200:00:09,019-->00:00:12,220感谢大家来到这场论坛Thanks,everyone.
Thanksforjoiningmeonthisdiscussionpanel.
300:00:12,814-->00:00:15,000很高兴今天能邀请到各位We'rehappytohaveyouguysonboardtoday.
400:00:15,063-->00:00:16,489不如先来自我介绍一下吧Howaboutaquickrunofintroduction500:00:16,490-->00:00:17,845从Malcolm开始Let'sstartwithMalcolm.
600:00:18,139-->00:00:20,556大家好我是Malcolm来自埃森哲Hi,I'mMalcolm,I'mfromAccenture.
700:00:20,557-->00:00:22,837负责技术战略业务LeaderofTechnologyStrategyPractice.
800:00:22,838-->00:00:23,895我在香港工作I'mbaseinHongKong900:00:23,896-->00:00:26,620但也负责大中华区的业务butlookaftertheGreaterChinabasesinregion.
1000:00:27,050-->00:00:29,310大家好我是JustinBairdHieveryone,mynameisJustinBaird,1100:00:29,311-->00:00:30,796也来自埃森哲I'malsoatAccenture,1200:00:30,797-->00:00:36,107是技术创新董事总经理I'mtheManagingDirectorofR&D,Technology&Innovation.
1300:00:36,843-->00:00:38,334我叫GerdLeonhard未来学家I'mGerdLeonhard,futurist,1400:00:38,335-->00:00:40,214来自瑞士苏黎世I'mbasedinZurich,Switzerland,1500:00:40,979-->00:00:43,850我的最新著作名叫《科技与人性》andIjustwroteabookcalledTechnologyVS.
Humanity,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
1600:00:43,881-->00:00:45,759其中谈到了人工智能whichtalksaboutAIaswell.
1700:00:46,398-->00:00:48,010我是Chwee来自IDCAndI'mChwee,I'mfromIDC,1800:00:48,011-->00:00:51,854人工智能与大数据分析Icoverartificialintelligenceandbigdataanalyticsglobally1900:00:52,043-->00:00:53,376是我的研究领域aspartofmyresearch.
2000:00:53,377-->00:00:55,654让我们开始采访吧Solet'sdiveintosomequestions,shallwe2100:00:55,655-->00:00:58,750当说到人工智能So,whenwementionartificialintelligenceorAI,2200:00:58,771-->00:01:02,066各位脑海中会联想到什么关键词吗isthereanysignificantsinglekeywordthatwillcometomind2300:01:02,067-->00:01:04,406对我来说是"天网"Malcolm你呢Formeit'sSkynet.
Malcolm2400:01:05,083-->00:01:07,195我的和你很不一样IthinkIwouldbequitefarfromSkynet,2500:01:07,196-->00:01:09,458我想到的是"人类"Iwouldsayit'sactuallyHuman.
2600:01:09,993-->00:01:13,241因为我认为人工智能的重点WhyhumanBecauseIthinkAIisactuallyallabout2700:01:13,242-->00:01:16,433是帮助人类以不同方式行事helpinghumandothingsdifferently2800:01:16,653-->00:01:19,743最终做到与众不同aswellasdoingdifferentthingseventually.
2900:01:19,744-->00:01:21,777所以我选择"人类"这个词So,Ichosethewordhuman.
3000:01:21,846-->00:01:24,154很有趣谢谢Justin呢Veryinteresting,thankyou.
JustinCopyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
3100:01:24,322-->00:01:27,704我认为人工智能Iactually,formeIthinkartificialintelligence3200:01:27,705-->00:01:30,318包括人工智能与自动化整个领域andreallykindofthewholeareaofAIandautomation3300:01:30,319-->00:01:33,179重点都在于赋能andthesethings,isallaboutenablement,3400:01:33,547-->00:01:35,416所以我的关键词是""赋能soEnablementtomeistheoneword.
Willitbettertotranslateenablementto赋能3500:01:35,417-->00:01:36,671因为它所关注的Becausethisisreallytalkingabout3600:01:36,672-->00:01:41,829是如何将人们从各种重复性工作中howwecantakepeoplefromrepetitivejobsandtasks3700:01:41,830-->00:01:43,986解放出来让他们远离此类工作andthesetypesofthings,andtakethemawayfromthat3800:01:43,987-->00:01:46,192去做更为有趣的事情andenablethemtodomoreinterestingthings3900:01:46,886-->00:01:49,334获得更多的工作满足感与生活幸福感andhavemorejobsatisfactionandhappierlife.
4000:01:50,562-->00:01:54,130我选择我在书中创造的一个词叫作"天狱"I'mgonnapickupawordthatImadeupinmybookcalledHelven,4100:01:54,172-->00:01:56,491即天堂和地狱whichisHellandHaven,right4200:01:57,402-->00:01:58,398我认为这个词很好Ithinkit'saverygoodword4300:01:58,399-->00:02:01,527因为人工智能可能成为天堂,也可能成为地狱becauseAIcouldbebothhellorcouldbeheaven.
4400:02:02,348-->00:02:04,728就像其他那些可靠的科技一样Likeanyotherreliabletechnology,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
4500:02:05,448-->00:02:07,681只有在为人所用时才是中立的it'snotneutraluntilweapplyit.
4600:02:08,282-->00:02:10,129所以我认为人工智能SoIthinkartificialintelligence4700:02:10,130-->00:02:13,907有着成就非凡的潜力hasthepotentialtoveryveryamazingthings,4800:02:13,908-->00:02:16,604也会有带来危害的潜力andhasotherpotentialswhicharenotsogood,4900:02:16,867-->00:02:21,561我们需要找到控制和引导它的方法we'llneedtofigureoutthewaytogovernitandtoguideit,5000:02:22,218-->00:02:24,247我认为这才是最主要的挑战andIthinkthatisthemainchallengereally.
5100:02:24,769-->00:02:28,032既然我们说到了机器问题Sincewearetalkingaboutthismachineryquestion,5200:02:28,033-->00:02:29,487那就来继续聊一聊let'sstayonthistopic.
5300:02:29,488-->00:02:31,506有人会很担心,比如说Istheregonnabeafearjustlike,youknow,5400:02:31,507-->00:02:34,287像源于英国的第一次工业革命那样thefirstindustrialrevolutioninEngland,5500:02:34,288-->00:02:38,324Justin也提到了willthatbeafearofthis,asJustinmentioned,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
5600:02:38,325-->00:02:40,706自动化会抢走我们的工作automationisgoingtotakeawayourjobs,5700:02:40,707-->00:02:43,705你认为我们该注意些什么呢whatshouldwebeaware,andwhatisyourthought5800:02:43,766-->00:02:44,530我认为首先Ithink,firstofall,5900:02:44,531-->00:02:47,067我们应该忘记所有看过的weshouldforgetallthe6000:02:47,068-->00:02:49,636那些好莱坞人工智能电影movieswe'veseenaboutAIfromHollywood,6100:02:49,637-->00:02:51,992包括印度和中国的(电影)那都不是现实fromIndiaorChina,itdoesn'treallymatter,6200:02:51,993-->00:02:53,660因为电影是为了娱乐becausethemoviesareaboutentertainment,6300:02:53,661-->00:02:57,247所以才会重点描绘两者间的战斗,等等whichmeansmostlyaboutwar,andthingslikethat.
6400:02:57,248-->00:02:59,970我们不该带着恐惧来看待人工智能Weshouldnotlookatthiswithfear,6500:03:00,645-->00:03:01,432同时atthesametime,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
6600:03:01,433-->00:03:06,130我们也不该一味幻想它就是好的wecannothavetheillusionthatthisisonlyjustgood,6700:03:06,339-->00:03:09,195与任何强大的技术一样likeanypowerfultechnology,6800:03:09,216-->00:03:11,539我们得找到折中点wehavetoactuallyfindthemiddlepath,6900:03:11,700-->00:03:15,061随着人工智能的到来很显然自动化Ithinkwhat'shappeningwithAIisquiteclearthatautomation,7000:03:15,275-->00:03:19,184机器人化与数字化会减少我们的很多工作robotizationandvirtualizationisgoingtokillalotofjobs,7100:03:19,240-->00:03:21,420非常多的工作manymanymanymanyofjobs,7200:03:21,808-->00:03:24,587同时也会创造出更多工作andotherjobswillbehappeningontheotherside,7300:03:24,758-->00:03:27,354有些人说在10年内somepeoplearesayingthatin10years7400:03:27,523-->00:03:31,189那些工作的70%都还没有出现about70%ofthosejobshavenotbeeninventedyet,7500:03:31,636-->00:03:33,390我认为这是好消息也是坏消息Ithinkthatisgoodnewsandbadnews,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
7600:03:33,391-->00:03:35,986但我们还是会进入这个生存状态butwearegoingtomovetoasociety7700:03:35,987-->00:03:39,419它里面的工作岗位与如今不同wherejobswillnotbewhattheyaretoday7800:03:40,076-->00:03:41,852不是为了生存likethepurposeofourlives,7900:03:42,396-->00:03:44,812那可能会是20年之后的事that'sprobablywillbe20yearsaway,8000:03:44,973-->00:03:47,111第二次改变我们的社会thendoublechangethesystem.
8100:03:48,712-->00:03:50,257Justin,你眼中的自动化Justin,toyourpointaboutautomation,8200:03:50,258-->00:03:52,623似乎更类似于一种扩张itseemslikeaugmentation,8300:03:52,840-->00:03:56,206人工智能可以提升我们的工作方法intelligenceempowerushowwedothings,howwework,8400:03:56,976-->00:03:59,157对很多程序实现自动化你是怎么看的automatealotofprocess.
Youropinion8500:03:59,458-->00:04:03,364我认为这一领域内已经有了些有趣的进展Ithinksomeinterestingworkhasbeendoneinthisarea,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
8600:04:03,584-->00:04:07,261人工智能可以大致分为10级wearekindoflookingatlet'ssayalmost10differentlevels,8700:04:07,451-->00:04:10,473最开始是近似型人工智能let'scallitstartingwithalmostAI,8800:04:10,533-->00:04:13,260它专注于简单的自动化任务reallyfocusonsimpleautomationandsometasks,8900:04:13,261-->00:04:14,613然后一直到allthewaythroughtothe9000:04:14,614-->00:04:16,570自动化的工作系统systemthatarekindofautomaticallyworking.
9100:04:17,193-->00:04:18,918当到了第10级So,wearelookingatlikelevel10,9200:04:18,919-->00:04:21,588比如说最为自动化的人工智能likethemostautonomouskindofAIstyle9300:04:22,030-->00:04:23,430它们很多都可以应用到andlotofthatcanbeappliedto9400:04:23,431-->00:04:25,861如今的企业运营中whatishappeninginanenterprisetoday.
9500:04:25,862-->00:04:28,353假设第10级的系统Sowe'llsayalevel10systemCopyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
9600:04:28,354-->00:04:31,807是可以从机器学习与训练中isonethatanalyzesthesituationfromwhatithaslearnt9700:04:31,808-->00:04:35,612习得环境分析技术的that'sgonethroughmachinelearningandit'satrainsystem.
9800:04:35,636-->00:04:37,511或许它面对一些事情时Anditmaybelookingatsomethings,9900:04:37,512-->00:04:39,994就可以自己作出决策andthenmakingthedecisionitself,10000:04:40,028-->00:04:41,335不需要人工andthentakingcareofthat10100:04:41,336-->00:04:43,948也不需要输入什么内容,就能自行解决问题withouthavinganybody,havinganyinputwhatsoever.
10200:04:43,949-->00:04:46,929所以人工智能可以从简单的Soitkindofrangesfrom--let'ssaysomethingmoresimple,10300:04:46,930-->00:04:48,164比如说第1或2级whichwouldbeatlevel1or2,10400:04:48,165-->00:04:50,148即询问"你想让我这么做吗"whichislike"doyouwantmedothisforyou"10500:04:50,149-->00:04:51,295一直到比如allthewaythroughtolikeyouknowCopyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
10600:04:51,296-->00:04:53,535你或许会在数据最后收到报告说youmightgetthereportattheendofthedatasays10700:04:53,566-->00:04:56,001"有25台服务器在运行"Therewere25serversrunning,10800:04:56,444-->00:04:58,200我关掉了4台4ofthemIturnedoff10900:04:58,201-->00:05:00,784因为它们没有执行程序只是在耗电"becausethereisnoprocessthatit'sjustburningmoney.
"11000:05:01,113-->00:05:03,277让你可以稍后进行更新Soyoujustgetupdatedlater.
11100:05:03,379-->00:05:06,761当说到人工智能在企业内的应用方式时WhenwethinkofthewaysAIcanbeusedinanenterprise,11200:05:06,762-->00:05:09,284其实不只是10年后I'mtalkingmore,notbythe10years,11300:05:09,285-->00:05:11,306人们在今天与之后数年内butmorelikepeopleareusingittoday11400:05:11,307-->00:05:12,355就已经在使用了andforthenextcoupleofyears,11500:05:12,356-->00:05:15,941这些来自各类人工智能系统中的优势thatkindofadvantagescomesfromthesetypesofsystemsCopyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
11600:05:15,942-->00:05:20,678我们可以称其为"自我思考"beingabletolet'scallit"Thinkingforitself".
11700:05:20,679-->00:05:22,463但是这种想法Butthatideaofactuallyhavingthat.
11800:05:22,464-->00:05:23,663我猜我们可以说Iguesswemightsay11900:05:23,664-->00:05:27,106如今企业内所使用的人工智能what'shappeningwithAIinanenterpriserightnowtoday,12000:05:27,188-->00:05:28,710不仅是.
.
.
morethankindofwhatwill--12100:05:28,711-->00:05:30,618它在未来也必定There'sdefinitelysignificantimpactsabout12200:05:30,619-->00:05:33,262对人类的工作环境有着显著影响howthataffectsontheworkplaceandeverythinginthefuture.
12300:05:33,810-->00:05:35,439我想再聊聊IfImightImayjustsayonething12400:05:35,440-->00:05:38,554自动化工作的敏感性aboutthesusceptibilityofjobsofautomation.
12500:05:38,555-->00:05:39,540这很有趣Ithinkit'squiteinterestingCopyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
12600:05:39,541-->00:05:42,536因为2013年有间不错的科研机构becausetherewasapinnaclestudyin201312700:05:42,537-->00:05:45,813做过这样一个尖端研究donebyproperscientificbodyresearch.
12800:05:45,814-->00:05:48,683他们收集了超过600个工作门类Theycategorizeover600jobcategories,12900:05:48,684-->00:05:51,504研究它们的敏感性与被自动化andtheylookatthesusceptibilityandprobabilityofthem13000:05:51,505-->00:05:53,995以及被替代的可能beingautomatedonacrossthatentirelist.
13100:05:53,996-->00:05:57,044那些极不可能被自动化替代的Sotheoneshasverylowsusceptibilityofautomation13200:05:57,045-->00:06:00,114包括考古学家物理治疗师withthinslikearchaeologists,physicaltherapists,13300:06:00,115-->00:06:02,403甚至是编舞evenlikechoreographers--13400:06:02,404-->00:06:06,490是需要用到很多不同认知能力的工作Thingsthattakealotofdifferenttypeofcognitiveability.
13500:06:06,607-->00:06:08,985然后再看另一边ThenifyoulookattheothersideofthelistofthingsCopyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
13600:06:08,986-->00:06:11,482很可能被替代的工作包括模特thatareverylikelytobeautomatedthingslikemodels.
13700:06:11,746-->00:06:17,114现在我们已经做出了惟妙惟肖的机器人Soalreadythere'sveryauthentichumanlookingrobotsoutthere,13800:06:17,329-->00:06:20,163比如大阪大学的GeminoidFlikeGeminoidFforexample,fromOsakaUniversity.
13900:06:20,164-->00:06:21,597还有我最看好的Butthenyoulookatmyfavorite14000:06:21,598-->00:06:24,099包括裁判以及其他体育赛事官员whichisumpires,refereesandothersportsofficials.
14100:06:24,100-->00:06:25,491所以他们基本上是说Sobasicallythey'resayingthat14200:06:25,492-->00:06:28,556"这个工作以后不会有了""Thisjobwillnotexistinthefuture.
"14300:06:28,557-->00:06:30,550你想想看网球比赛时Andifyouthinkaboutwatchingthetennis14400:06:30,825-->00:06:35,117场地会更清爽没有裁判没有边裁it'sjustfetter,theumpire,thelinejudgers,it'sjustfetter.
14500:06:35,118-->00:06:37,861因为如果他们不相信裁判的判定Becauseiftheydon'ttrustwhatthepersonsaid,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
14600:06:37,862-->00:06:39,523他们会相信电脑theytrustthecomputer.
14700:06:39,524-->00:06:41,743他们会把镜头放到最大看回放Theysay"Oh,showme"theyzoomallthewayin,14800:06:41,744-->00:06:44,525没人会想过电脑的判定不准确noonewillexpectthatthecomputerwilleverbeoutofalignment,14900:06:44,526-->00:06:47,130他们会依此判罚theyjusttrustthatthisiswhattocallit.
15000:06:47,131-->00:06:49,482我们能在不久的将来看到这些不同的事It'sinterestingtoseehowthesedifferentthings15100:06:49,483-->00:06:52,352影响我们的生活很有意思willbeimpactineventhenearfuture.
15200:06:52,558-->00:06:55,424Malcolm,Justin提到了企业层面的Malcolm,Justinmentionedabouttheenterprise-based,15300:06:55,425-->00:06:56,704人工智能应用AIenterprisedbased.
15400:06:56,705-->00:07:00,530那么埃森哲对此有何策略又是如何What'sAccenture'sstrategyandhowareyouguyslookingat15500:07:00,531-->00:07:04,017帮助企业运用这一新技术的呢helpingenterprisesutilizingthisnewtechnologyCopyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
15600:07:04,938-->00:07:06,600其实在我们最近的调查里Infact,inourrecentresearch,15700:07:06,601-->00:07:09,617我们发现相比两年前wefoundthatcomparedto2yearsago,15800:07:10,212-->00:07:13,183无论是大型还是小型企业,对此的投入theinvestmenthasgrownupmorethan70%,15900:07:13,366-->00:07:15,537都增长了70%多bycorporations,largeorsmall.
16000:07:15,538-->00:07:19,799人工智能已经得到应用并且会留存下来AIisactuallyheretodayandit'sgoingtostay.
16100:07:19,997-->00:07:25,859我用个简单的等式来说明人工智能的价值IfImayputasimpleequationtoarticulatethevalueofAI,16200:07:25,860-->00:07:30,675第一个因素应该是人类andthisiswhereIthinkthefirstelementisaroundthepeople.
16300:07:31,459-->00:07:34,590如果没有人类人类不去努力工作ifwedon'thavethepeople,peopledon'tperformthejobwell,16400:07:34,596-->00:07:38,075不去创造、创新就不会有人工智能peopledon'tcreate,innovate,IthinktherewillnotbeAI.
16500:07:38,640-->00:07:40,412第二个是流程Thesecondpartistheprocess,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
16600:07:40,413-->00:07:42,525围绕着我们所说的自动化过程processaroundtheautomationwetalkabout,16700:07:42,526-->00:07:43,797围绕着新的商业模型aroundthenewbusinessmodel16800:07:43,798-->00:07:46,332它可以被技术创造出来thatcanbecreatedofthetechnology,16900:07:46,536-->00:07:48,504第三个就是数据andthethirdoneisaboutthedata.
17000:07:48,996-->00:07:53,971所以人类加流程加数据,就是这个公式的核心Solookatthis,thepeople,thehuman,theprocess,thedata,17100:07:54,036-->00:07:55,379我们可能会有疑问wesaysowhat17200:07:55,380-->00:07:58,677这三者存在挺长时间了为什么他们代表着未来Wehavetheseforquitesometime,that'swhywecallfuture,17300:07:58,864-->00:08:02,395有趣的是,人工智能是随之出现的Well,TheinterestingthingisAIcomesalong,17400:08:02,522-->00:08:04,755你不仅有着这三种因素whereyoudon'tjusthavethesethreeelements,17500:08:04,756-->00:08:06,593现在你有了科技butnowyouhavetechnology.
Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
17600:08:06,714-->00:08:10,931它不仅能理解行为还能学习Don'sjustsensethecomprehendact,youcanalsolearn.
17700:08:11,264-->00:08:14,052所以说到底人工智能Sotheyearout,AIisreallyabout17800:08:14,053-->00:08:16,310是将这三种因素结合bringingthesethreeelementstogether17900:08:16,651-->00:08:20,519自己以指数级的力量位居顶端andyouputtheexponentialpowerofAIontopofthat,18000:08:20,520-->00:08:22,871让你可以真正放大thisiswhereyoucanreallymagnify18100:08:22,872-->00:08:25,841人类流程和数据的影响力theimpactofthepeople,processanddata.
18200:08:25,842-->00:08:29,116拿企业作为例子来说Sotakingorganizationasanexample,18300:08:29,117-->00:08:31,453很多投资公司youknowmanyinvestmentfirms18400:08:31,454-->00:08:34,543做算法交易已经很久了aredoingalgorithmictradingforalonglongtime,18500:08:34,544-->00:08:37,507很多企业都在提倡虚拟代理Manycompaniesaretrollingvirtualagent,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
18600:08:37,508-->00:08:39,461我的很多客户都已经用上了manyofmyclientsarealreadyhavingthese.
18700:08:39,462-->00:08:41,552他们会在聊天窗口内Theclientstalkingtothechatbox18800:08:41,553-->00:08:44,397询问保险政策获取建议byinsurancepolicy,andtakingadvice.
18900:08:44,398-->00:08:46,718所以这些如今都已存在Soalltheseareheretoday,19000:08:46,729-->00:08:50,234你的孩子们在和Alexa与Siri互动yourkidsareplayingandtalkingtoAlexa,Siri,19100:08:50,235-->00:08:51,197这都是真实的sothisisreal.
19200:08:51,198-->00:08:54,741而且我们相信.
.
.
包括其他设备Andwedobelievethatthishasa--andinstallations,19300:08:54,892-->00:08:58,816这对企业会有着数百亿元的影响we'retalkingabouttensofbillionsofimpacttotheenterprise,19400:08:58,817-->00:09:00,921而且我觉得它会留存下来andIthinkit'sgoingtostay,19500:09:00,922-->00:09:05,643我知道企业们会愿意接受人工智能IwouldsuggestIknowtheenterpriseiswillingtoembracethisCopyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
19600:09:06,061-->00:09:08,998并且对其托付信任andtakealeapoffaithforwardonthesubject.
19700:09:08,999-->00:09:10,866是的谢谢你MalcolmRight,thankyou,Malcolm.
19800:09:10,867-->00:09:12,231下一个问题是给Justin的NextquestionforJustin.
19900:09:13,076-->00:09:14,965我们最近看到了很多并购We'veseenalotofacquisitionsrecently,20000:09:15,173-->00:09:18,376谷歌买下了很多开源平台Google'sbuyingoutalotoftheopensourcekindsoffloor,20100:09:18,377-->00:09:21,703很多科技公司包括英特尔也是they'rebuyingalotoftechnologycompanies,evenIntel,right20200:09:22,684-->00:09:24,390我们发现Wekindofseeingthese20300:09:24,391-->00:09:26,407这些并购变得更为垄断acquisitionshavebecomemoremonopolizing,20400:09:26,408-->00:09:29,080最新最尖端的人工智能公司一出现LatersandgreatersAIcompaniestostartup,theycomeup,20500:09:29,081-->00:09:31,014立刻就被收购andtheygetboughtoutveryquickly.
Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
20600:09:31,017-->00:09:33,613很多技术也被移向云端Andalotofthesehavebeenmovedtothecloud,too.
20700:09:34,127-->00:09:36,290你认为这是趋势还是机遇呢Youseethisasatrendoristhisanopportunity20800:09:36,291-->00:09:37,978我们需要担心这种情况吗Isthissomethingweshouldbeconcernedabout20900:09:37,979-->00:09:41,389人工智能的知识产权WheresomuchconsolidationoftheseAIIPshavebeenconsolidated21000:09:41,390-->00:09:44,939被大公司收购Malcolm你怎么看atthesebigmajorcorporationsMalcolm21100:09:45,684-->00:09:47,018我觉得这很有意思Ithinkit'sinteresting.
21200:09:47,019-->00:09:50,661大型的平台企业突然间都掀起了Youlookatthetrendandsuddenlythebigplatformplayers21300:09:51,128-->00:09:54,983收购人工智能技术的浪潮acquiringAItechnologiesatarapidspeed.
21400:09:54,984-->00:09:59,660当你看到未来的成功包括如今的Andyoulookatthefuturesuccesses,eventoday,21500:09:59,661-->00:10:02,913当人们提到人工智能会想到谁whenpeopletalkaboutAI,whoareyouthinkingaboutCopyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
21600:10:03,515-->00:10:05,577人们会想到谷歌PeoplewillactuallythoughtaboutGoogle,21700:10:05,578-->00:10:08,282会想到亚马逊peoplewillthoughtaboutAmazon,forexample.
21800:10:08,283-->00:10:11,004这就让大家开始思考Andthiskindofmakespeoplestartthinking21900:10:11,065-->00:10:16,293为什么他们能在人工智能方面如此成功呢whyarethesebigplayerscanbesosuccessfulinAI22000:10:16,434-->00:10:18,123其实这并不在于人工智能It'sactuallynotaboutAI.
22100:10:18,140-->00:10:20,816我们相信数据才是重点Wedobelieveit'sactuallyaboutdata.
22200:10:20,891-->00:10:25,584如果没有数据来训练人工智能机器Becauseifyoudon'thavedatatotrainyourAImachine,22300:10:25,917-->00:10:28,254它就不可能实现智能youactuallywon'tgetsmarter.
22400:10:28,291-->00:10:33,044所以你才会看到这些技术方案Thisiswhyyouseethekindoftechnologyandsolutions22500:10:33,045-->00:10:34,917包括数据开始不断融合andthedataisactuallyconverging,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
22600:10:34,918-->00:10:36,078这也是为什么andthisiswhereweseethe22700:10:36,079-->00:10:39,002大型平台公司会有市场largeplatformplayeractuallyhasaniche.
22800:10:39,003-->00:10:41,803因为他们有数据可以训练机器Becausetheyhavethedata,theycantrainthemachine.
22900:10:42,079-->00:10:46,372想想无人驾驶汽车想想特斯拉Youlookatthedriverlesscar,youlookatwhatTeslaisdoing,23000:10:46,373-->00:10:49,317它的核心都是在训练算法it'sallabouttrainingthealgorithms.
23100:10:49,318-->00:10:52,197比如谷歌的模型NowyouknowtheGoogle'smodel,forexample,23200:10:52,198-->00:10:55,601他们可以完全免费theycanmakeitallfree,allfree.
23300:10:55,602-->00:10:58,662就像他们让安卓平台免费一样JustlikehowtheymakeAndroidfreetothepeople.
23400:10:58,749-->00:11:01,352我们能让谷歌的人工智能WearegoingtohavethisGoogleAI23500:11:01,353-->00:11:03,333在人们的口袋里运行runninginprobablypeople'spocket,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
23600:11:03,336-->00:11:09,772帮助他们来训练加强这些算法helpingthemtrain,andenhancethesealgorithmsandgetsbetter.
23700:11:09,773-->00:11:11,323我觉得这样的模型Ithinkit'squiteadifferentmodel23800:11:11,324-->00:11:14,395与亚马逊和Azure是不同的fromhowAmazonandAzureispurging.
23900:11:14,396-->00:11:16,024后者的也很好Ithinkthisisallfine.
24000:11:16,025-->00:11:20,239但说实话我认为在将来ButIthinkthefuture,ifIputmymindonthetable,24100:11:20,240-->00:11:22,504这些大平台企业Ithinkthislargeplatformplayer24200:11:22,505-->00:11:25,921才会真正地功成名就willactuallysucceedandglorified.
24300:11:26,437-->00:11:27,834Justin你认为呢Justin,whatdoyouthink24400:11:27,963-->00:11:32,654我同意我认为在自由的市场环境内Iagree,Ithinkinafreemarketenvironment.
24500:11:33,943-->00:11:37,042我们可能之前也聊到了Ithinkas--wemighthavediscussedbefore,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
24600:11:37,043-->00:11:38,909我在谷歌工作了差不多6年I'veworkedforGoogleforabout6years,24700:11:38,910-->00:11:41,521看过很多不同的并购Isawanumberofdifferentacquisitions.
24800:11:41,522-->00:11:42,775我认为这很有意思It'sinteresting,Ithink.
24900:11:42,846-->00:11:45,444像你说的这会存在很多机遇There'salotofopportunitiesintermsofasyouaresaying,25000:11:45,445-->00:11:49,804你能获得一些独特的机会youaretakingsomeuniqueopportunities,25100:11:50,709-->00:11:54,598获得比现有条件更多的数据gainalotmoreoutofperhapsthedatayoualreadyhave,right25200:11:54,599-->00:11:57,033我们在很多领域中都见过这样的情况Ithinkweseethathappenacrossdifferentareas.
25300:11:57,709-->00:12:01,835比如在埃森哲我们有图像识别系统Forexample,inAccenturewehaveanimagerecognitionsystem25400:12:01,836-->00:12:06,671可以用来进行商业与科研活动thathasbeenusedforbothcommercialandtechnology.
25500:12:07,351-->00:12:12,135很有趣的一点是大多数问题Andwhat'squiteinterestingaboutthatismostoftheproblems,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
25600:12:12,136-->00:12:14,322比如图像识别系统如何工作orlikehowanimagerecognitionsystemworks,25700:12:14,323-->00:12:17,406如何学习或训练这些大家都知道andhowitlearnedandhowit'strained,thesethingsareknown.
25800:12:18,267-->00:12:20,457你可以研究开放计算机主义Youcanlookatstufflikeopencomputerism,25900:12:21,568-->00:12:23,446如果你想识别出照片或视频中的人像mostofthethingsthatyouwanttodoifyouwantto26000:12:23,447-->00:12:25,616你需要做大部分事情recognizeafaceinthepictureorinthevideowhatever,26100:12:25,617-->00:12:26,605都是运用已知的技术thosethingsareknown.
26200:12:26,606-->00:12:29,680而知识产权的核心来自于其训练方式WhatisthecoreIPcomesfrom,howithasbeentrained,26300:12:29,681-->00:12:30,855训练平台andwhatit'sbeentrainedon,26400:12:30,856-->00:12:32,905以及通过系统训练带来的价值andthevaluethatsystemcandeliverfromthattraining.
26500:12:32,906-->00:12:34,045所以像你说的Soasyousay,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
26600:12:34,046-->00:12:36,893我们把这些因素结合创造出一种webringthattrainingtogethertocreatekindoflikea26700:12:39,178-->00:12:41,577集体大于个体的形式thesumofthepartisgreaterthenthewholeheart.
26800:12:41,697-->00:12:43,118我能再补充两句吗CanIjoinhimjusttomakeitmoreclear26900:12:43,119-->00:12:45,312我觉得人们必须意识到Ithinkit'simportanttorealize27000:12:45,313-->00:12:46,988我们在说的大部分事情thatmostofthethingswearetalkingaboutthere27100:12:46,989-->00:12:51,337其实是关于系统智能化的isreallyaboutIA,intelligentthesystems,27200:12:51,404-->00:12:54,691我们现在看到的成果中的98%that's98%ofallthegoodstuffthatweareseeingnow,27300:12:54,692-->00:12:57,424对很多企业来说都很容易实现that'sthelow-hangingfruitformanycompanies.
27400:12:57,816-->00:13:00,938包括提高地图准确度提高客户关系管理That'sstufflikemakingthemapsbetter,orhavingbetterCRM,27500:13:01,058-->00:13:03,985人脸识别声音识别recognizingpeople'sfaces,recognizethevoice,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
27600:13:03,986-->00:13:05,990这都是智能方面的事that'salltheintelligentstuff,27700:13:06,032-->00:13:08,737和能思考的机器的关系微乎其微thathasverylittletodowiththinkingmachines.
27800:13:09,465-->00:13:11,395如果真有能思考的机器的话Ifthereiseveractuallysuchathing.
27900:13:11,396-->00:13:13,027因为人工智能的下一级BecausethenextlevelofAI28000:13:13,028-->00:13:17,686也就意味着让机器自己创造东西whichmeanssomethingthathappensthemachinekindofcreates.
28100:13:18,720-->00:13:21,671我们在这方面看到的不多We'veseenalittlebitofthatbutnotmuch,28200:13:21,672-->00:13:24,903主要是花哨的高端软件之类mostoftheotherstuffarebasicallylikefancysoftware,28300:13:25,152-->00:13:26,758那算是智能thatiskindofintelligent.
28400:13:26,759-->00:13:29,746我认为这种情况还要持续几年Ithinkthat'sgonnabelikethisforafewyears.
28500:13:29,940-->00:13:31,865然后下一步才能到人工智能AndthenthenextstepisartificialintelligenceCopyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
28600:13:31,866-->00:13:34,234然后再是普遍智能andthenafter,generalintelligence,28700:13:34,389-->00:13:37,234比如说,检测心率变异的机器whichwillbeforexamplethemachinesthatplaysHRV28800:13:37,235-->00:13:41,957也能指挥交通等等canthenalsoruntrafficcontrol,thatsortofthing.
28900:13:42,119-->00:13:45,199那才是最不同寻常的点Thatisthepointofsingularity29000:13:45,200-->00:13:48,005当机器可以做出人类可轻易做到的行为iswhenthemachinecanactuallydothingsthatweeasilydo.
29100:13:48,006-->00:13:51,188比如你会开车Forexample,youcanhavetheskilllikedriveacar,29200:13:51,214-->00:13:55,411那你也能轻易地驾驶摩托艇等等thenyouwouldeasilyapplyittodriveamotorboatorsomething,29300:13:55,525-->00:13:57,116这点机器永远办不到thatmachinecanneverdo,29400:13:57,880-->00:13:59,649这是完全不同的Andthat'scompletelydifferent.
29500:13:59,650-->00:14:01,917我认为很多人担心人工智能IthinkalotofpeopleareworriedaboutAIalso,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
29600:14:01,918-->00:14:03,652是因为他们觉得becausetheythinkthat29700:14:03,831-->00:14:06,022智能机器的思考方式和我们一样thinkingmachinearethinkactuallylikewedo–29800:14:06,023-->00:14:07,933其实并不是andthatisjustnotthecase.
29900:14:08,560-->00:14:11,248这是关键区别之一Thisisoneofthekeydifferentiate,30000:14:11,249-->00:14:13,321即对于现今的企业来说isthatforcompaniesandforbusinessestoday30100:14:13,322-->00:14:15,657人工智能还只是智能系统it'saboutintelligencesystems,period.
30200:14:16,135-->00:14:19,644不管它是什么这才是首要的Nomatterwhatitis–thatisthefirstbigthing,youknow30300:14:20,606-->00:14:24,780Justin我想继续你刚才说到的那点Justin,Iwantyoutocarryonthatpointyoujustmentioned.
30400:14:25,126-->00:14:26,743它的潜在使用Potentialusageforthat.
30500:14:26,744-->00:14:28,303Justin你也提到了AndJustinyou'vementionedtoo,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
30600:14:28,543-->00:14:31,776埃森哲也在做很多计算机视觉研究等等Accenturearedoingalotofresearchestoo,computervisionetc.
30700:14:31,777-->00:14:35,938你能分享一些埃森哲所做的研究AreyouabletosharesomeoftheresearchthatAccenturedo30800:14:35,939-->00:14:39,274以及它可以如何应用于企业吗andhowitcouldbeappliedtotheenterpriseside30900:14:41,525-->00:14:43,781好作为埃森哲的部门Yes,aspartofAccenture,31000:14:43,782-->00:14:48,989我们研发部在进行几项不同的研究we'vegotafewofdifferentthingsintermsofR&D,31100:14:48,990-->00:14:52,049我们或许该重点说说传统研发maybeweshouldstarttalkingmoreaboutclassicalR&D,31200:14:53,079-->00:14:56,072埃森哲在全球都有研发技术实验室wehaveR&Dtechlabsaroundtheglobe,31300:14:56,073-->00:14:59,079这些实验室关注的是在未来5到10年内andthoseR&Dtechlabsarereallygonnafocusonthetechnology31400:14:59,080-->00:15:02,486将会对商业包括对环境thataregonnabeimpactinthebusinessesandourenvironment31500:15:02,947-->00:15:04,344产生影响的科技inthenext5to10years.
Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
31600:15:04,345-->00:15:05,924他们在提前做着准备Theyaredoingkindofworkearlier,31700:15:05,925-->00:15:07,876正处于实验阶段kindofstageontheexperimentation.
31800:15:08,790-->00:15:12,325研究涉及了很多领域Thetypeofworkdoesrangemanyareas.
31900:15:13,045-->00:15:17,772我们将这些研发的部分Oneofthekindof,fromtherewetaketheseR&Dcomponents32000:15:17,773-->00:15:20,396以及有形的机遇andoneofthosetangibleopportunities,32100:15:20,634-->00:15:23,088带入到下一次的机遇中wecanbringthoseintokindofnextopportunities32200:15:23,089-->00:15:25,987用于诸如流动工作室等等whichisIguesswheregroupsliketheliquidstudios,32300:15:25,988-->00:15:27,287它是在某种领域whichisoneofthenew32400:15:27,288-->00:15:30,407以及世界范围内的新方法practicesofaroundtheregionandtheworld,32500:15:30,469-->00:15:33,317这个想法比如说已被证实的概念takethatideaandsayhereistheprovedconcept,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
32600:15:33,318-->00:15:34,570感兴趣的客户hereistheclientthatinterestedin32700:15:34,571-->00:15:35,988与这样的技术结合workingwiththisstepoftechnology,32800:15:35,989-->00:15:40,265用这些想法来创造有形的产品taketheseideasandcreatearealtangibleproduct,32900:15:40,307-->00:15:42,105为客户带来机遇andopportunityforthatclient.
33000:15:42,106-->00:15:45,665进入他们的系统站在他们的视角.
.
.
Sofitintotheirsystems,workingwiththeirkindofposition--33100:15:47,313-->00:15:49,806我们在不同领域都有研发ImeanthereisIguessworkingindifferentareas,33200:15:49,807-->00:15:50,665其中一个andoneoftheareasis33300:15:50,666-->00:15:53,962主要就是机器人与人工智能领域thereisalotoffocusonroboticsandartificialintelligence.
33400:15:53,963-->00:15:59,605我们认为协助是人工智能未来的趋势之一Andtheassistantareaisthestuffweseeiskindofcoming.
33500:16:00,597-->00:16:03,907比如说智能顾问这个概念Soforexamplethereisaconceptofsmartadvisors.
Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
33600:16:03,908-->00:16:06,238你可以把智能顾问Smartadvisorsaresomethingforexample33700:16:06,882-->00:16:09,324看作在你电脑后台运行的一段程序that'srunninginthebackgroundonyourcomputer,33800:16:09,325-->00:16:12,501比起Siri或微软小娜andinsteadofSiriorCortana33900:16:12,502-->00:16:15,405或者其他你需要激活的助手orsomethingthatyougoandactionitandsay"HeySiri.
"34000:16:15,519-->00:16:17,025这个智能顾问正相反Thisiskindoftheoppositeofit,34100:16:17,026-->00:16:21,437系统会主动和你来交流whichisthesystemgoes"Heyyou","HeyJustin".
34200:16:21,438-->00:16:23,842它会出现说出问题Soitcomesupandsaysoneofthethings34300:16:23,843-->00:16:26,112比如你正在研发新软件likewhenyou'redevelopingsoftware,34400:16:26,729-->00:16:29,374在向智囊团介绍它youintroducethesoftwaretotherepository,34500:16:29,375-->00:16:31,455这时候智能系统就可以出现说that'swhenthesmartsystemscomesupandsaysCopyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
34600:16:31,456-->00:16:33,973"让我们来看看你做的成果如何""OK,let'sseewhatyou'vedone.
"Anditmightgo--34700:16:34,143-->00:16:36,454或许你的软件里可能存在句法错误等等Maybethere'ssomesyntaxthingsorwhateveritmightbe,34800:16:36,455-->00:16:37,674它就会弹出来说anditwillpopupandsay34900:16:37,675-->00:16:39,248"你刚才遗漏了一段代码"Hey,thecodeyou'vejustomitted35000:16:39,249-->00:16:40,723你可以这样修复"there'sacoupleofthingsyoucanfix.
"35100:16:40,724-->00:16:42,328我可以问"还有谁在编写这段代码"AndthenIcansay"Whoelseisworkingonthiscode"35200:16:42,329-->00:16:44,128它就会告诉我其他还有正在编写的人Itwouldtellmethedifferentpeopleworkingonthat.
35300:16:44,129-->00:16:46,309就好像你知道这些.
.
.
Thereislikeyouknowthesekindofthings【这段声音太小】35400:16:48,578-->00:16:51,570所以它像是更先一步的建议Soit'smorelikeaproactive,pushrecommendations.
35500:16:51,571-->00:16:53,527对它能看到你正在做的事Yeah,soit'skindoflearn--it'sseeingwhatyou'redoingCopyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
35600:16:53,528-->00:16:55,070并在整个过程中帮助你andassistingyouintheprocess.
35700:16:55,071-->00:16:55,788是Right.
35800:16:56,525-->00:16:59,361而且它不可以说"你这样很蠢"Andit'sprogrammedtonotsay"Hey,thatwasstupid.
"35900:17:02,556-->00:17:03,382Malcolm你有什么看法Malcolm,anycomments36000:17:03,383-->00:17:08,256我们的很多研究关注的是人工智能应用IthinkalotofresearchfocusisactuallyfromtheappliedAI,36100:17:08,354-->00:17:11,608所以我们没有发明AlphaGosowearenotthecompanythatwillinventAlphaGo,36200:17:11,722-->00:17:13,094也不像谷歌wearenotthecompanythat36300:17:13,095-->00:17:16,633或者亚马逊那种平台开发公司wanttodeveloptheplatformlikeGoogleandAmazon.
36400:17:16,774-->00:17:19,858我们所做的,就是理解科技Butwhatwedoisweunderstandtechnology,36500:17:20,055-->00:17:23,993因为我们非常贴近客户行业与市场becauseweareveryclosetoourclients,industryandbusiness,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
36600:17:23,994-->00:17:26,994了解商业难题了解客户特性weknowthebusinessproblem,weknowwhatthecustomers,36700:17:27,346-->00:17:29,538可以将两者结合weareabletocombinethetwotogether.
36800:17:29,539-->00:17:31,702这才是我们很多科技实验室Thisiswherealotofourtechlab,36900:17:31,703-->00:17:35,179流动工作室与创新实验室所关注的liquidstudio,innovationlabfocusedon.
37000:17:35,180-->00:17:36,999而且我们有不同的层级比如Andwehavedifferenttiers,forexample,37100:17:37,000-->00:17:41,119从技术孵化器到科技实验室fromtheincubationfrontadventurerightdowntotechlab,37200:17:41,121-->00:17:43,743到创新实验室到流动工作室toourinnovationlab,andtoliquidstudio.
37300:17:43,744-->00:17:46,404从不同层级不同的科技成熟度Sodifferentlevel,maturityoftechnologies,37400:17:46,538-->00:17:49,542到不同的应用范围todifferentscaleofapplicability,37500:17:49,739-->00:17:53,448到全球客户这就是我们研究的关注方式toourclientsglobally,that'showourresearchesarefocused.
Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
37600:17:53,449-->00:17:55,755是如果错了请纠正Right.
AndcorrectmeifI'mwrong,37700:17:55,756-->00:17:58,388因为你们横跨各个行业becauseyouguysworkingacrossallverticals,37800:17:58,478-->00:18:00,698任何行业都有埃森哲的身影soanykindofverticalsyou'llbethere,37900:18:00,711-->00:18:02,837在寻找解决问题的方法lookingathowwecansolvethisproblem.
38000:18:03,306-->00:18:06,005没错拿人工智能来说Absolutely.
TakeAIasanexample,38100:18:06,013-->00:18:07,613我们的研究实际上ourresearchisactually38200:18:07,614-->00:18:11,725分布在40个子行业内dividedacross40sub-industriesverticals.
38300:18:12,700-->00:18:16,879当然它们都是可以应用的常见科技Ofcoursetheyareallcommontechnologythatcanbeapplicable,38400:18:16,891-->00:18:19,514但我们向客户提供的解决方案butthesolutionsthatweareadvisingourclients38500:18:19,515-->00:18:22,743实际上是很有行业针对性的isactuallyquiteindustry-specific.
Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
38600:18:23,062-->00:18:26,655因为40个子行业内的使用环境都不同Becausetheusecasesaredifferentacross40sub-industries.
38700:18:27,486-->00:18:30,729当你们每次说起人工智能时Whatisthebiggestquestionyougetfromyourcustomers38800:18:30,730-->00:18:32,701听到客户最大的疑问是什么Everytimeyoustarttalkingaboutthistopic38900:18:32,702-->00:18:35,014他们会问你现在是人工智能的严冬They'regonnaaskyouwhetherit'sanotherwinterAI39000:18:35,015-->00:18:37,089还是布局的最佳时机orthisreallythetimetojumpin.
39100:18:38,924-->00:18:40,470从我个人的经验来看Justfrommypersonalexperience,39200:18:40,471-->00:18:43,882我认为人们知道人工智能已经来到我们身边Ithinkpeopleare--theyknowthisishappening,39300:18:43,883-->00:18:47,049也知道自己需要跟上这一创新的节奏andtheyknowthattheyneedtokeepupwiththisinnovation.
39400:18:47,581-->00:18:49,474但很多时候人们都想搞明白Butalotoftimespeoplearetryingtofigureout39500:18:49,475-->00:18:51,063应用它的最佳方式thebestwaytoapplythat.
Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
39600:18:51,064-->00:18:51,871我认为很多问题Ithinkalotofthis39700:18:51,872-->00:18:55,239的确是来自人们听说了很多人工智能doescomefrompeoplehearingalotaboutAI,39800:18:55,727-->00:18:57,892但或许他们并不理解butthenkindof--maybetheyarenotunderstanding39900:18:57,893-->00:19:01,036人工智能如何运用到他们的经验中howAIcanbebroughtintotheirexperience--40000:19:02,637-->00:19:05,374比如说一个很好的例子Soforexample,thegoodonejustkindof…40100:19:06,937-->00:19:11,221现在银行也有了语音互动应用That'sforexample,banksnowcomingupwithAlexaappandstuff40200:19:11,222-->00:19:14,527你可以说"Alexa我的账户余额如何"youcansay"Hey,Alexa,what'smybalanceandstatments"40300:19:15,403-->00:19:17,321这些是可以查到的对吧There'sthosetypeofthings,right40400:19:18,161-->00:19:19,082有趣的是What'sinterestingthereis40500:19:19,083-->00:19:21,916这种不用动手的查询设计很棒it'sgreattohavethatwhereit'skindofahands-free,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
40600:19:21,917-->00:19:26,060让人们可以轻装上阵.
.
.
it'skindofdesignforhumansratherthanhavingtouseyour--40700:19:28,188-->00:19:31,090但人们也需要体验Butthenit'sliketheirneedstobetherightexperience--40800:19:32,616-->00:19:36,208我知道对我自己包括很多人都会问WhenIknowformyself,andIthinklotsofpeoplewouldsay40900:19:36,209-->00:19:37,537"我的账户明细如何""Oh,what'smystatementlistthere"41000:19:37,538-->00:19:39,816你能说是或不是这就成了对话Youcansayyesorno,it'skindofaconversation.
41100:19:39,817-->00:19:42,126而不仅仅是说"您的余额如下"Insteadofsaying"Oh,hereisyourstatement.
"41200:19:42,127-->00:19:45,126然后系统开始报出一串数字Thensystemstartsramblingoffabunchofnumbers.
41300:19:45,198-->00:19:45,878但你真正想要的Butreallywhatyouwant41400:19:45,879-->00:19:47,890是一个随之显示的仪表板isadashboardthatcomesalongwiththat.
41500:19:48,256-->00:19:50,059所以你还是想有这种警卫在身边Soit'slikeyouwanttohavethekindofexperienceCopyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
41600:19:50,060-->00:19:52,625人工智能帮助你查询账户的实际体验whereyouareguardedandtheAI'shelpingyoudothesethings.
41700:19:52,626-->00:19:55,437但你也可以随之获得Butthenyouactuallygetdifferenttypesofexperience41800:19:55,438-->00:19:56,387不同的体验togoalongwiththat.
41900:19:56,388-->00:20:00,809获得视觉互动感受我们在努力创造的Getthatkindofvisual--feellikethoseend-to-endexperience42000:20:00,810-->00:20:02,444那些点对点的体验thatwe'retryingtofocuson.
42100:20:03,820-->00:20:04,572而不是光说Notjustsaying42200:20:04,573-->00:20:07,522"对是人工智能用吧""Ohyes,AI,thisthing,let'splugin.
"42300:20:07,913-->00:20:09,828我们得弄清该带来怎样的体验Wegottofigureouttherightexperiencestobring.
42400:20:09,829-->00:20:11,715是Malcolm呢Right,gotit.
Malcolm42500:20:11,716-->00:20:13,459我同意因为我认为Iagree,becauseIthinkCopyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
42600:20:13,460-->00:20:16,837我们的客户关于人工智能是什么there'sactuallyalotofconfusionamongstourclients,42700:20:17,579-->00:20:18,657还是有很多疑惑的whatisAIabout.
42800:20:18,658-->00:20:23,237比如他们以为人工智能就是买IBMWatsonTheythoughtAIisaboutbuyingsomeIBMWatson,forexample.
42900:20:23,245-->00:20:27,130就是买亚马逊的一些人工智能服务Theythoughtit'saboutbuyingsomeAIservicesfromAmazon.
43000:20:27,989-->00:20:29,439其实并不是No,it'snot.
Right43100:20:29,471-->00:20:32,666你得从消费者的角度去思考Yougottothinkfromacustomer'sperspectives.
43200:20:32,667-->00:20:34,261你想解决什么问题Whatproblemsyouwanttosolve43300:20:34,262-->00:20:36,735你想给人们带来怎样的体验Whatexperienceyouwanttodelivertopeople,right43400:20:36,893-->00:20:39,153然后你得考虑企业中的因素Andyoulookatwhatyouhaveinenterprise43500:20:39,161-->00:20:44,103包括你的人员流程数据thatincludeyourpeople,yourprocess,yourdata,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
43600:20:44,104-->00:20:48,460人工智能的美妙之处在于它是虚拟代理andthenthiswonderfulthingonAIisit'savirtualagent.
43700:20:48,461-->00:20:52,131它有机器学习能力等等There'ssomelearning,machinelearningcapabilityandsoforth.
43800:20:52,132-->00:20:53,677然后你将它们结合Andyoubringthesetogether.
43900:20:53,678-->00:20:55,622但这需要时间Butit'sactuallyajourney,44000:20:55,623-->00:20:58,327因为没有数据你无法实现人工智能becauseyoucannotjumptoAIwithouthavingthedata,44100:20:58,369-->00:21:00,945你的组织里withoutsomeofthesedigitalDNA44200:21:00,946-->00:21:02,840没有这些数字基因sortitoutinyourorganization.
44300:21:02,870-->00:21:04,406所以我会建议客户SoIwouldadviseourclients44400:21:04,407-->00:21:06,448去整体地看待人工智能的应用you'vegottolookatthisholistically,44500:21:06,466-->00:21:10,326从小处开始然后慢慢扩大startsmall,andthentobeabletoscaleupgradually.
Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
44600:21:10,352-->00:21:11,921好感谢MalcolmGotit,thanksMalcolm.
44700:21:12,177-->00:21:13,637我们还剩几分钟Sowehaveaboutcoupleofminutesleft,44800:21:13,638-->00:21:14,579我想来做个总结Ijustwanttowrapitup.
44900:21:14,580-->00:21:16,721也算是一种引导和建议吧Soit'skindofaguidance,good.
45000:21:17,081-->00:21:18,834对于我们普通人来说Forusnormalpeople,right45100:21:18,835-->00:21:21,390不是大企业你们有什么建议吗Notbigenterprises.
What'syourguidanceonthis45200:21:21,391-->00:21:26,321各位都是顶尖创新者通晓这一领域Imean,you'releadinnovator,knowalotofstuffofthisarea.
45300:21:26,322-->00:21:30,195那么对我们这些普通人有什么指导吗What'syourguidanceforusnormalfolksinthisarea45400:21:30,619-->00:21:32,481首先我想说IthinkfirstIwouldsaythat45500:21:32,482-->00:21:37,996我们既要谨慎也要保持好奇心weneedtobebothcarefulbutalsocurious.
Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
45600:21:38,455-->00:21:40,292我们不该说Weshouldnotgoandsay,45700:21:40,293-->00:21:41,708"到时候会出现这些问题"Ok,there'sgonnabealltheseproblems45800:21:41,709-->00:21:43,164我才不要参与呢"andIwouldn'tgetinvolved.
"45900:21:43,165-->00:21:45,166但同时我们也不该因为自己有能力Butatthesametime,weshouldn'tjustdoanything46000:21:45,167-->00:21:46,473就为所欲为justbecausewecan.
46100:21:47,173-->00:21:48,244我认为最重要的Ithinkmostimportantly,46200:21:48,245-->00:21:51,306是人们要意识到人的独特性Ithinkpeoplearerealizinghowuniquepeopleare,46300:21:51,307-->00:21:52,828人类的独特性howuniquehumansare.
46400:21:53,060-->00:21:55,824我们不该用科技来少这种独特性Weshouldnotusetechnologytomakethatsmaller46500:21:55,825-->00:21:57,088而是使其增加buttomakethatbigger.
Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
46600:21:57,340-->00:22:02,608所以我们既要小心也要主动Soit'sbasicallyamixtureofbeingprecautionaryandproactive.
46700:22:02,609-->00:22:03,339是Right.
46800:22:04,269-->00:22:05,670因为这一科技Becausethisisalsothetechnology46900:22:05,671-->00:22:07,698可以彻底地改变世界thatcouldchangetheworldentirely47000:22:08,241-->00:22:10,767就算不是今后5年内也是20年内notinthenext5years,butin20years.
47100:22:11,096-->00:22:14,309因为这一科技我们或许Andweprobablyneedtochangeourentirestructure47200:22:14,310-->00:22:18,105也需要改变我们的整个工作与合作模式ofhowweworkandhowwecollaboratebecauseofthistechnology.
47300:22:18,325-->00:22:21,810我觉得最关键的Ithinkthecrucialthingisreallyabout47400:22:22,282-->00:22:24,686是该让政治家和普通人理解gettingpoliticiansandpeopletounderstand47500:22:24,687-->00:22:26,154人工智能对社会意味着什么whatthatmeansforthesociety.
Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
47600:22:26,546-->00:22:27,888我认为它主要是有益的Ithinkit'smostlygood,47700:22:27,889-->00:22:30,956但也可能有很多需要说明的副作用butitcouldhavemanysideeffectsthatweneedtoaddress.
47800:22:30,957-->00:22:32,966没错MalcolmJustinTotallyagree.
Malcolm,Justin,47900:22:32,967-->00:22:35,488你们俩所说的主要是sowhatIheardfrombothofyouisbasically48000:22:35,543-->00:22:37,362埃森哲所提供的whatAccenturebringtotableis48100:22:37,363-->00:22:40,402是一种透明度youactuallyprovidesthetransparency,48200:22:40,768-->00:22:43,504让企业去了解什么是人工智能providesenterprisestounderstandwhatAI'sabout48300:22:43,505-->00:22:46,042以及它能如何帮助企业解决问题andhowitactuallyhelpedthemsolvetheirproblem48400:22:46,043-->00:22:48,535这就是你们所创造的最佳商业价值andthat'swhatthebestbusinessvalueyouguysmade.
48500:22:48,536-->00:22:51,454作为总结你们还有什么想说的吗AnyothercommentsthatyoucanprovidetowrapthisupCopyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
48600:22:52,256-->00:22:52,896好Sure.
48700:22:52,897-->00:22:56,343比如我们想要帮助客户Forexample,wethinkabouthelpingourclients48800:22:56,344-->00:22:59,397在这条创新之路上继续前进movealongkindofthisapathofinnovation.
48900:22:59,577-->00:23:00,629我认为很多时候Ithinkalotoftimes49000:23:00,630-->00:23:02,873人们都在说人工智能.
.
.
peopleuseoneisthatAIisgetting--49100:23:03,323-->00:23:05,411就像创新一样justasmuchasinnovationwere.
49200:23:05,412-->00:23:06,425这究竟是什么意思Likewhatdoesthatevenmean49300:23:06,426-->00:23:08,009大家怎么定义这个词呢Howdoesanyonekindofdefinethat49400:23:08,021-->00:23:11,064对我来说创新是有价值的改变Sotome,innovationislikechangethathasvalue.
49500:23:11,065-->00:23:14,636任何有价值的改变都是创新它可以简单Soanychangethathasvalue,itcouldbesomethingsimple,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
49600:23:14,642-->00:23:17,397我最喜欢的一个老派例子就是likeoneofmyfavoriteoldschoolexampleisthat49700:23:17,398-->00:23:19,611星巴克曾经有过这个点子平台Starbucksusedtohaveanideasplatform,49800:23:19,612-->00:23:21,248任何人都能提出自己的构想anyonecanjustputtheirideasin,49900:23:21,249-->00:23:23,811一位是威斯康辛还是密苏里的女士IthinkwomanfromlikeWisconsin,orMissouri,50000:23:23,812-->00:23:26,296我记不太清了有一天说Icannotrememberexactlywhere,onedaysaid50100:23:26,297-->00:23:27,568"我喜欢你们的冰咖啡"Iloveyouricecoffees,50200:23:27,569-->00:23:29,441但每次我喝到杯底时butatthetimeIgottothebottomtheicecoffee50300:23:29,442-->00:23:30,072都只剩水了it'sjustwater,50400:23:30,073-->00:23:32,153你们为什么不直接用咖啡做冰块呢"sowhydon'tyoujustmakeicecubesoutofcoffee"50500:23:32,474-->00:23:35,241这是个很简单的例子很好的创新Whatagreatsimpleexample,greatinnovation,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
50600:23:35,242-->00:23:37,559他们在全世界推行现在冰咖啡变得更好喝theyrolledthatoutglobally,sotheicecoffeeisbetternow.
50700:23:37,921-->00:23:39,363它可以是这种小问题Itcouldbesomethingreallysimplelikethat,50800:23:39,364-->00:23:41,632也可以是需要使用人工智能oritcouldbeusingartificialintelligence50900:23:41,633-->00:23:43,514帮助解决的世界性问题tohelpsolveworldproblems.
51000:23:44,477-->00:23:46,533我们目前参与的一项任务Oneofthethingsweareinvolvedwithrightnow51100:23:46,534-->00:23:51,738是用人工智能协助生存在全世界isusingAItohelpclassifyallofthecoralandfish51200:23:51,739-->00:23:53,717礁石区的鱼类与珊瑚的分类thatarelivingonreefsaroundtheworld.
51300:23:53,836-->00:23:56,314我们使用人工智能图像识别系统Sowe'reusinganAIimagerecognitionsystem51400:23:56,315-->00:23:57,992来进行分类togoandclassifythesethings.
51500:23:58,018-->00:24:00,057这样能监控它们的情况Sowecanactuallykeeptrackonwhat'shappeningwiththat,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
51600:24:00,058-->00:24:01,309参与这个participatinginthenumber51700:24:01,310-->00:24:03,237名为"50礁"的全球性研究计划ofglobalprojectsonecalledthe50Reefs,51800:24:04,153-->00:24:06,586帮助保护50片lookingtohelpkeepthe50reefs51900:24:06,587-->00:24:11,048最能适应海洋复杂性thataremostresilienttooceansophistications,52000:24:11,191-->00:24:13,291气候变化气温与种群变化的礁石climatechanges,youknow,temperature,race,52100:24:13,292-->00:24:15,537我们帮助他们做这些分类andwe'rehelpingthemactuallydothiskindofclassification,52200:24:15,538-->00:24:16,377当统计结束sowhentheygoback,52300:24:16,378-->00:24:18,775他们就能看到鱼类或珊瑚数量theycanseewhetherfishpopulationsorcoral52400:24:18,776-->00:24:20,901是增加还是减少了haveincreasedordecreasedinpopulation.
52500:24:21,153-->00:24:22,752如果用人工智能进行分类AndifyouareusingAItodothis,Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
52600:24:22,753-->00:24:24,407它的速度it'slike900timesfaster52700:24:24,408-->00:24:27,047会比让大学里的学生进行thanhavingresearchstudentsattheuniversity52800:24:27,048-->00:24:28,586逐张手工分类快900倍goingthrougheverysinglepicture.
52900:24:28,587-->00:24:29,744而且是全世界一百多所大学We'retalkinghundredsofuniversityaroundtheworld.
53000:24:29,745-->00:24:30,624是Right.
53100:24:31,171-->00:24:34,576所以使用人工智能造福人类非常棒SousingAItobenefitisagreatthing,53200:24:34,707-->00:24:36,498但我也同意你说的butItotallyagreewithyouaswell.
53300:24:36,766-->00:24:39,699对于人工智能在消费者与企业中的推广Gottabeverycarefulabouthowthesethingsaredeployed53400:24:39,700-->00:24:41,980也需要非常谨慎bothtoconsumersandtoenterprises.
53500:24:41,981-->00:24:43,101好谢谢你JustinRight,thankyouJustin.
Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
53600:24:43,102-->00:24:44,000Malcolm最后补充几句吧Malcolm,finalwords53700:24:44,001-->00:24:46,890好我再多说一点Yeah,somaybeIthinkjustonepoint.
53800:24:46,891-->00:24:48,590我们需要边界Becauseweneedaterritory,53900:24:48,721-->00:24:53,398总有一天我们会进入普遍学习时代thatwecouldbeenteringintothisgenerallearningoneday,54000:24:53,399-->00:24:57,799这是很好的机会但也很危险thisisspotopportunitybutit'salsoquitedangerous.
54100:24:57,878-->00:24:59,721对埃森哲这样的企业来说SoforcompanieslikeAccenture,54200:24:59,722-->00:25:02,549在为客户推荐方案时wegottobeveryresponsible,54300:25:02,550-->00:25:06,168我们就需要非常地负责intermsofsolutionthatwearerecommendingtoourclients.
54400:25:06,371-->00:25:08,214机器的学习能力Theabilityforthemachinetolearn.
54500:25:08,215-->00:25:11,910我们得在道德范围内进行操作Wegottooperatewithinthisethicalboundary.
Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
54600:25:12,041-->00:25:15,502所以在埃森哲我们称其为负责任的人工智能SoinAccenture,wecallthisresponsibleAI.
54700:25:15,577-->00:25:16,914如果做不到这点Ifwedon'thavethat,54800:25:16,915-->00:25:22,074而是贪心地发挥出人工智能的全部威力andjustgoalongwiththemaximumpower,fullthrottleofAI,54900:25:22,078-->00:25:24,454我认为我们就会进入Ithinkwecouldbegettinginaterritory55000:25:24,455-->00:25:26,921可能做出不道德行为的领域thatwecouldbedoingunethicalthings.
55100:25:26,922-->00:25:29,607-或者产生反作用力-反作用力没错-Orbacklash.
-Orbacklash,exactly.
55200:25:31,343-->00:25:33,917我们要负起责任这点很重要Weareresponsible,isactuallyquiteimportant,55300:25:33,918-->00:25:37,800对于埃森哲这样的企业来说甚至是空前重要的evenmoreimportantthanbeforeforfirmslikeAccenture.
55400:25:38,196-->00:25:40,003好感谢大家今天的到场Okay,thankyoueveryoneforjoiningustoday.
55500:25:40,004-->00:25:42,059我想感谢MalcolmJustin与GerdI'dliketothankMalcolm,JustinandGerdCopyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.
55600:25:42,060-->00:25:43,421参与这次论坛tobeonthisdiscussionpanel.
55700:25:43,422-->00:25:45,362感谢各位精彩的见解Greatinsightsandthankyouverymuch.
55800:25:45,363-->00:25:47,650-谢谢-谢谢-Thankyou.
-Thankyou.
Copyright2017AccentureAllrightsreserved.
Accenture,itslogo,andHighPerformanceDeliveredaretrademarksofAccenture.