观世界 | 全球知名企业高管预测2019人工智能趋势

20 More AI Predictions For 2019

数据观|和璟祎（编译）

Artificial intelligence (AI) is everywhere, driven by large investments, lots of startups, all established technology vendors, and enterprises big and small experimenting with what it can do for their bottom line.

人工智能(AI)无处不在，在大型投资、大量初创公司、所有成熟的技术供应商以及大大小小的企业的推动下，他们都在试验AI能为他们的利润做些什么。

“Some AI Applications will not live up to the hype, and that's OK. People have been planning to have self-driving cars for a while. Some still fear an AI take over might be just 20 years away but the truth is we're still a long way away from truly autonomous cars. And as for an AI takeover, that will only exist in SciFi movies for the foreseeable future. My prediction is that our expectations for AI and the reality of its capability will meet somewhere in the middle. The next 5 years will look a lot like they do now, but our day-to-day will become more and more efficient in subtle, yet significant, ways. AI bots will get better at answering questions and vetting customer service cases, smart assistants will be more equipped to complete tasks and self-driving car features will continue to improve, but they will not take over the road”—Richard Socher, Chief Scientist, Salesforce

“一些人工智能应用会辜负炒作，但这没关系。人们计划拥有自动驾驶汽车已经有一段时间了。一些人仍然担心人工智能可能在20年后将会接管，但事实是，我们距离真正的自动驾驶汽车还有很长的路要走。至于人工智能的收购，在可预见的未来，这只会出现在科幻电影中。我的预测是，我们对人工智能的期望和它的能力的现实将在中间的某个地方达到。未来5年看起来和现在很像，但我们的日常生活将变得越来越有效率，以微妙但重要的方式。人工智能机器人将在回答问题和审查客户服务案例方面变得更好，智能助手将拥有更多完成任务的设备，自动驾驶汽车的功能将继续改进，但它们不会取代道路”

——Salesforce首席科学家Richard Socher。

“The adoption of artificially intelligent offerings will continue to scale into different verticals from manufacturing to education, retail and more in 2019. In the healthcare sector, for example, AI-enhanced applications have the capability to reduce emergency waiting room times and even free up doctors’ time through the use of AI in detecting and diagnosing tumors. As new advances and applications make their way into various verticals, expect to see accelerated adoption as technology costs come down and organizational and business outcomes improve. At Lenovo, we’re already using AI in our supply chain and parts planning process so that we can better develop best in class experience for customers also keen to transform their business with artificial intelligence”—Gianfranco Lanci, Corporate President and Chief Operating Officer, Lenovo

“2019年，人工智能产品的应用将从制造业继续扩展到教育、零售等各个垂直领域。例如，在医疗行业，人工智能增强的应用程序能够通过使用人工智能来检测和诊断肿瘤，减少急诊候诊室的时间，甚至解放医生的时间。随着新的科学进展和应用程序进入各个垂直领域，技术成本的降低与组织和业务结果的改善，我们预计这将加速人工智能的采用度。在联想，我们已经在供应链和零部件规划过程中使用了人工智能，以便更好地为那些同样渴望使用人工智能改造业务的客户提供一流的体验”

——联想公司总裁兼首席运营官Gianfranco Lanci。

“Patients will find themselves talking via a variety of omni-channel UIs in addition to the pre-existing chat bots that are currently available on mobile apps and other health care IT platforms. Consumer frameworks for conversational experiences like Alexa and Google Home may add HIPAA privacy support that opens the gates for bots to keep the dialog going during the big gaps in time between patient visits. In care settings, nurse call buttons beside hospital beds, forms to collect health histories, and klunky scheduling apps will evolve into customer-focused robot medical assistants”—Dan Housman, chief technology officer for ConvergeHEALTH, Deloitte

“患者会发现，除了现有的移动端APP和其他互联网医疗平台的聊天机器人之外，他们还可以通过各种全通道用户界面进行医患交谈。像Alexa和Google Home这样的会话体验，用户框架或许可以添加HIPAA（《健康保险流通与责任法案》）隐私支持，这将为机器人打开一扇大门，以便在患者在前后访问的间隔时间内保持适时沟通。在护理机构，医院病床旁的护士呼叫按钮、病历表格以及残旧的排程应用程序，都将发展成为以客户为中心的机器人医疗助手。”

—— 德勤(Deloitte)总经理兼德勤医疗系统项目ConvergeHEALTH首席技术官Dan Housman。

“2019 will see the pendulum shift to a focus on performing analytics at the edge. Organizations will save time and money by processing and analyzing data at the edge versus moving it back to a core, storing it and applying traditional analytics. Use cases include anomaly detection (fraud), pattern recognition (predicting failures/maintenance) and persistent streams. Autonomous vehicles, Oil and gas platforms, medical devices are all early examples of this trend that we will see expand in 2019. Cost drivers for this trend are bandwidth (semi-connected environments as well as expensive cellular) considerations and storage (reduce the amount of data sent to the cloud)”—Jack Norris, Senior Vice President, Data and Applications, MapR

“2019年我们将看到钟摆式的转折——聚焦边缘分析的执行。与其传统地将数据移回核心存储、应用，企业更青睐于边缘处理、分析数据以节省时间和成本。其用例包括异常检测(欺诈)、模式识别(故障预测/维护)和持久流。自动驾驶汽车、油气平台、医疗设备都是这一趋势的早期例子，我们将在2019年看到这一趋势的扩展。这一趋势的成本驱动因素是带宽(半连接环境以及昂贵的蜂窝网络)和存储(减少发送到云端的数据量)”

——开源大数据技术公司MapR数据与应用程序高级副总裁Jack Norris。

“Public demands for responsible AI will increase. 2018 was the year of awakening. 2019 will be the year of action. It won't be just data ethicists and human rights advocates demanding fairness, accountability, and transparency. Consumers are already changing how they use Facebook or deleting their accounts altogether and this trend is likely to spread to other social media and other services that leverage personal data. Greater numbers of pledges and declarations about the responsible creation and use of AI will be written and companies will be pressured to adopt them. The public will fight back over government use of biased AI in decisions impacting human rights. More employees will demand influence over what they create and refuse to contribute to harmful automation. Companies will have to lead with their conscious-- whether they are buying AI solutions or building them-- and seek assurances that the systems are fair in order to avoid being the next headline on AI gone awry”—Kathy Baxter, Architect of Ethical AI Practice, Salesforce

“公众对可靠人工智能的需求将会增加。如果说2018年是AI觉醒之年，那2019年就将成为行动之年。不仅是数据伦理学家和人权倡导者要求公平、问责和透明，消费者已经在改变他们使用Facebook的方式，或者干脆删除他们的账户，这种趋势可能会蔓延到其他通过注册个人数据信息来登录的社交媒体和其他服务。更多关于有责创造和使用人工智能的承诺书及声明将形成书面文件，企业将被迫采用遵循，公众也将对政府在影响人权的决策中使用有偏见的人工智能进行反击。更多员工将要求对他们所创造的东西施加影响，并拒绝为有害的自动化做出贡献。企业将不得不以自己的意识为先导——无论是购买人工智能解决方案还是建立人工智能，都要确保这些系统是合法的，以避免成为下一个‘问题AI’头条。”

——Salesforce伦理AI实践架构师Kathy Baxter。

“Advanced analytics and artificial intelligence will continue becoming more highly focused and purpose-built for specific needs, and these capabilities will increasingly be embedded in management tools. This much-anticipated capability will simplify IT operations, improve infrastructure and application robustness, and lower overall costs. Along with this trend, AI and analytics will become embedded in high availability and disaster recovery solutions, as well as cloud service provider offerings to improve service levels. With the ability to quickly, automatically and accurately understand issues and diagnose problems across complex configurations, the reliability, and thus the availability, of critical services delivered from the cloud will vastly improve”—Jerry Melnick, President and CEO, SIOS Technology

“高级分析和人工智能将继续变得更加集中，并为特定的需求而专门构建，这些功能将越来越多地嵌入管理 工具 中。这个备受期待的功能将简化IT操作，改进基础设施和应用程序稳健性，并降低总体成本。随着这一趋势的发展，人工智能和分析技术将被嵌入到高可用性和故障恢复解决方案中，以及云服务提供商提供的提高服务水平的产品中。由于能够快速、自动和精确地理解问题，跨越复杂配置诊断问题，云端关键服务的可靠性和可用性将极大提高。”

——SIOS Technology总裁兼首席执行官Jerry Melnick。

“As chatbots and AI continue to evolve, the depth and breadth of functions they can perform will increase. What does this mean for the workforce, positively and negatively? On one hand, machine learning will help people sift through massive amounts of data and do their jobs more effectively. On the other, customer service and support roles will be phased out as people grow more comfortable with bot interactions. This will begin to occur on a wider scale in 2019, as more enterprises adopt AI and chatbots to either boost productivity among their existing workforce, or phase out positions that can be taken with the assistance of these technologies”—David Cohn, Co-founder and Chief Strategy Officer, Pigeon

“随着聊天机器人和人工智能的不断发展，它们能够执行的功能深度和广度将会增加。这对劳动力意味着什么，积极还是消极?一方面，机器学习将帮助人们筛查大量数据，并更有效地完成工作。另一方面，随着人们对机器人交互越来越熟悉，客户服务和售后角色将逐步消失。到2019年，随着越来越多的企业将采用人工智能和聊天机器人来提高现有员工的生产率，或者逐步淘汰这些技术可以替代的职位，这种情况将大范围出现。”

—— Pigeon联合创始人兼首席战略官David Cohn。

“A dirty little secret about industrial-strength AI is that many of these systems are trained and evaluated on datasets created and labeled by thousands (or more) human raters. As we tackle more complex AI problems, the need for massive amounts of high-quality human judgments will increase, but there will be breakthroughs in leveraging machine learning techniques to make collecting those judgments more time- and cost-efficient. At the same time, methods which use minimal or even no labeled data (aka unsupervised techniques) will reduce our reliance on large swaths of labeled data, enabling deep learning models to be more robust on new and different types of problems”

—Joel Tetreault, Head of Research, Grammarly

“关于工业级人工智能的一个不可告人之处是，这些系统多数是由成千上万甚至更多人类评分者创建和标记的数据集进行培训和评估的。随着我们解决更复杂的人工智能问题，对大量高质人类判断的需求将会增加，但在利用机器学习省时省钱收集这些判断方面将会出现突破。与此同时，使用最少甚至不使用标记数据的方法(又称无监督技术)将减少我们对大量标记数据的依赖，从而使深度学习模型能够更稳健地处理新的、不同类型的问题。”

——Grammarly研究主管Joel Tetreault。

“Knowledge Graphs are the new black! The technologies needed – NLP, Graph DB, Content Analytics – are now aligned to enable knowledge graphs to easily codify domain knowledge. From usable chatbot, guided processes to automated advisors, we’ll see increased use in many industries and domains, including healthcare, financial services, and supply chain”—Jean-Luc Chatelain, Managing Director & Chief Technology Officer, Accenture Applied Intelligence

“知识图是新的黑马!所需的技术——NLP、图形数据库、内容分析——现在已看齐以使知识图能够轻松地编码领域知识。从可用的聊天机器人、引导流程到自动化顾问，我们将看到他们在众多行业和领域里的使用率增加，包括医疗保健、金融服务和供应链。”

——Accenture Applied Intelligence总经理兼首席技术官Jean-Luc Chatelain。

“AI has moved into the mainstream with innovations in self-driving cars, smart speakers, and facial recognition. Less visible, yet equally impactful, are AI applications around logistics, manufacturing, healthcare, and cybersecurity.  And what makes cybersecurity unique is that it’s an essential component of all the other technologies. Whether we choose to live in an ‘intelligent’ or an ‘artificially intelligent’ world, one thing is certain: If AI and deep learning isn’t enhancing your cybersecurity strategy, you’re far more likely to get hacked. AI makes it considerably more difficult for cybercriminals to earn their disreputable income.  With an AI-powered defense, attackers are left to seek out softer targets (those who don’t think they need AI) or they’re forced to develop even more sophisticated and costly methods of attack – and so the cyber arms race continues”—Joe Levy, CTO, Sophos

“随着自动驾驶汽车、智能扬声器和面部识别技术的创新，人工智能已进入主流。虽然人工智能在物流、制造、医疗保健和网络安全方面的应用不那么引人注目，但同样具有影响力。网络安全的独特之处在于它是所有其他技术的重要组成部分。无论我们选择生活在一个“智能”还是“人工智能”的世界，有一件事是肯定的:如果人工智能和深度学习不能增强你的网络安全战略，你就更有可能被黑客攻击。人工智能让网络犯罪分子更难获得声名狼藉的收入。有了人工智能的防御，攻击者只能寻找更容易的目标(那些认为自己不需要人工智能的目标)，或者被迫开发出更复杂、成本更高的攻击方法——因此，网络军备竞赛仍在继续。”

——Sophos首席技术官Joe Levy。

“AI is entering the Age of Commodity. You do not need to know how the technology of a microwave works in order to use it, it is simply a tool. With the huge influx of no-code, point-and-click tools we are entering into the same phase with AI where it will become a widely used utility by everyone, regardless of technical background. As a result, most of the AI applications in the coming years will be built by people with little or no AI training”—Vitaly Gordon, VP Data Science, Salesforce

“人工智能正在进入商品时代。你不需要知道微波技术如何工作才能使用它，它只是一个工具。随着大量无代码、点击式工具的涌入，我们正进入与人工智能相同的阶段，在这个阶段，无论技术背景如何，它都将成为一种广泛使用的实用工具。因此，未来几年的大部分人工智能应用程序将由很少或没有人工智能培训的人开发。”

——Salesforce数据科学副总裁Vitaly Gordon。

“Robotic Process Automation (RPA) has been one of the hottest areas of tech in the last two years – because of its simple, easy-to-understand value prop – process automation, efficiency; freeing resources up to focus on higher value activities, etc. But It has fundamental limits – it’s only effective with rote, repetitive processes and it cannot impact workflows involving unstructured content which makes up over 80% of data in most enterprises. At the same time, AI and machine learning are seen as too esoteric; requiring too much data science expertise, too much hand-holding, too much uncertainty and risk about ROI. Companies will look to bridge the gap in 2019 – between the horsepower of RPA and the intellect of AI/machine learning through what many experts are calling ‘intelligent process automation”—Tom Wilde, CEO and Founder, Indico

“机器人流程自动化(RPA)在过去两年中一直是最热门的技术领域之一——因为其简单、易于理解的价值主张——流程自动化、效率高，释放资源以专注于更有价值的活动等等。但它有基本的局限性——它只对死记硬背、重复的流程有效，不会影响涉及非结构化内容的工作流，在大多数企业中，非结构化内容占数据的80%以上。与此同时，人工智能和机器学习被视为过于深奥，需要太多的数据科学专业知识、太多扶持、太多关于利润率的不确定性和风险。2019年，各公司有望通过专家们提出的“智能流程自动化”，在RPA马力和人工智能/机器学习之间架起一座桥梁。”

—— Indico首席执行官和创始人Tom Wilde。

“Artificial Intelligence (AI) and machine learning (ML) are overhyped for many real-life applications, including the contact center industry. For example, instead of trying to identify specific patterns in images or data (an AI/ML sweet spot), it will be much more useful to increase the volume of satisfying self-service support sessions through intelligently applied automation to resolve common questions and provide guided user flows through defined business processes. By leveraging human intelligence primarily for those support scenarios that can’t be effectively automated, call center operations will be further optimized”—Anand Janefalkar, Founder and CEO, UJET

“人工智能(AI)和机器学习(ML)在许多实际应用中被夸大了，包括呼叫中心行业。例如，与其试图识别图像或数据中的特定模式(AI/ML的最佳点)，不如通过智能应用自动化来解决常见问题并通过定义的业务流程提供指导用户流程，从而增加满足自助服务支持会话的数量。通过主要利用人类智能来支持那些不能有效自动化的场景，呼叫中心的操作将得到进一步优化。”

—— UJET创始人和首席执行官Anand Janefalkar。

“In 2019, there will be a shift from AI toolkits to AI solving specific enterprise challenges, such as IT and human resources employee experiences. To date, the model has been that enterprises can apply hard-to-come-by skills to leverage AI toolkits to build a custom application. This is shifting to using AI to solve common enterprise problems”—Pat Calhoun, Founder and CEO, Espressive

“2019年，将会有一个从人工智能工具包到人工智能解决具体企业挑战的转变。如IT和人力资源员工的经验，迄今为止，模型成了企业利用AI工具包构建自定义应用程的难获技术。这正转向使用AI来解决常见的企业问题。”

——Espressive创始人和首席执行官Pat Calhoun。

“In 2019, AI’s early adopters in the enterprise will look to gain more value from their AI investments as they expect more abundant and richer, built-in AI solutions within cloud applications in terms of functionality, user experience, and accessibility (multi-device, chatbots, digital assistants, etc.).  We’ll see companies investing in third party data sources and smart data (dynamic signals and flexible classifications that are regularly refreshed) to optimize outputs. Trust, transparency, and explainable AI will become bigger issues as organizations wrestle with machine learning bias. Customers will realize that machine learning requires human supervision and features like supervisory controls, coupled with data insights, to help early adopters tweak machine learning outputs and generate more value from AI investments”—Melissa Boxer, VP of Adaptive Intelligent Applications, Oracle

“2019年，人工智能在企业中的早期采用者希望从他们的人工智能投资中获得更多价值，因为他们希望云应用程序中内置的人工智能解决方案在功能、用户体验和可访问性(多设备、聊天机器人、数字助理等)方面更丰富。我们将看到公司投资于第三方数据源和智能数据(动态信号和定期刷新的灵活分类)以优化输出。随着企业与机器学习偏差作斗争，信任、透明度和可解释的人工智能将成为更大的问题。客户将会意识到机器学习需要人类的监督和监督控制等功能，再加上数据洞察，以帮助早期采用者调整机器学习输出，并从人工智能投资中产生更多价值”

—— Oracle自适应智能应用副总裁Melissa Boxer。

“Marketers have long talked about taking the ideal next best action when it comes to marketing programs based on where people are in the buying cycle. However, this has been impossible to achieve without massive amounts of data being synthesized by AI in real time. The emergence of AI to take over manual tasks involving huge data sets means that automated next best action triggered by specific activity in the buying cycle will become a reality in 2019”—Peter Isaacson, CMO, Demandbase

“长期以来，营销人员一直在谈论，当涉及到基于人们在购买周期中所处位置的营销计划时，应该采取仅次于最佳的理想行动。然而，如果没有人工智能实时合成的海量数据，这是不可能实现的。人工智能的出现接管了涉及庞大数据集的手工任务，这意味着在2019年，由购买周期中的特定活动触发的自动次优行动将成为现实。”

——Demandbase首席营销官Peter Isaacson。

“In 2019, incorporating AI will be an essential part of the marketing strategy. Trained models around predictive analytics, sentiment analysis, programmatic advertising, to name a few, will revolutionize how marketers automate more aspects of the marketing pipeline and develop highly targeted Account Based Marketing (ABM) strategies. This will require investment in new technologies but will also lower custom acquisition costs by making marketing dollars more effective”—Daniel Raskin, CMO, Kinetica

“2019年，整合人工智能将成为营销战略的重要组成部分。围绕预测分析、情绪分析、程序化广告等方面，训练有素的模型，将彻底变革营销人员在营销渠道上多方面的自动化，并开发出具有高度针对性，基于客户的营销(ABM)策略。这需要对新技术进行投资，但也将通过提高营销资金的有效性来降低定制采购的成本。”

——Kinetica首席营销官Daniel Raskin。

“Artificial intelligence and machine learning will become a requirement for new solutions for simplified operations: The IT skills gap will require progressive enterprises to implement new, innovative solutions that automate complex operations. Machine learning and artificial intelligence will become key requirements for new IT solutions to help businesses close the skills gap through smarter operations and modern IT solutions. Enterprise software firms will force their strategic vendors to integrate AI and ML into their existing offerings to provide a more efficient operating model and a higher level of success for meeting their desired outcomes”—Don Foster, Senior Director, Commvault

“人工智能和机器学习将成为简化操作新解决方案的需求:IT技能的差距会导致先进企业实现自动化复杂操作的创新解决方案。机器学习和人工智能将成为新的IT解决方案的关键需求，以帮助企业通过更智能的操作和现代IT解决方案缩小技术差距。企业软件公司将迫使他们的战略供应商将人工智能和机器学习集成到他们现有的产品中，以提供更有效的操作模型和更高层次的成功来满足他们期望的结果。”

——Commvault资深总监Don Foster。

“Nearly every IT department will adopt AI to automate enterprise monitoring, reduce manual work of IT staff, and enable a vision of applications that can repair themselves”—Dave Anderson, Digital Performance Expert, Dynatrace

“几乎每个IT部门都将采用人工智能来实现自动化企业监控，减少IT人员的手上工作，并实现使应用程序能够自我修复的愿景。”

—— Dynatrace数字显示专家Dave Anderson。

“The robotics industry will see many startups trying to find a niche and trying to capture as much market share possible. However, in order to succeed, robotics startups must consider regulatory regulations from the start of designs so that they meet applicable safety regulations or else they will fail when they go to market”. Ryan Braman, Test Engineering Manager, TUV Rheinland

“我们在机器人行业将会看到许多初创公司试图找到一个利基市场，并尽可能多地抢占市场份额。然而，为了取得成功，机器人初创企业必须从设计的一开始就考虑监管规定，以满足适用的安全法规，否则在进入市场时就会失败。”

——TUV Rheinland测试工程经理Ryan Braman。

注：《全球知名企业高管预测2019人工智能趋势》来源于 Forbes (点击查看原文)。数据观编译/和璟祎，转载请注明译者和来源。

责任编辑：李兰松