David R. Clough (david.clough@sauder.ubc.ca) is an assistant professor in the OBHR Division and the Entrepreneurship and Innovation Group at the University of British Columbia’s Sauder School of Business. He received his PhD from INSEAD. His research interests include new venture emergence, organizational learning, and technological change in innovation ecosystems. David R. Clough (david.clough@sauder.ubc.ca) is an assistant professor in the OBHR Division and the Entrepreneurship and Innovation Group at the University of British Columbia’s Sauder School of Business. He received his PhD from INSEAD. His research interests include new venture emergence, organizational learning, and technological change in innovation ecosystems.
Andy Wu (awu@hbs.edu) is an assistant professor in the Strategy Unit at Harvard Business School. He received a PhD and MS in applied economics from the Wharton School of the University of Pennsylvania, where he is a senior fellow at the Mack Institute for Innovation Management. 吴安迪（awu@hbs.edu）是哈佛商学院战略研究部的助理教授。他拥有宾夕法尼亚大学沃顿商学院应用经济学博士和硕士学位，目前在沃顿商学院的创新管理麦克因斯研究所担任高级研究员。
Data Network Effects: Key Conditions, Shared Data, and the Data Value Duality
数据网络效应：关键条件、共享数据与数据价值二元性
Clough and Wu (2022) provide an interesting and thought-provoking response to our article (Gregory, Henfridsson, Kaganer, & Kyriakou, 2021) on the role of artificial intelligence (AI) and data network effects for the creation of user value. We welcome the debate around data network effects as a new category of network effects. In this response note, we build upon the points raised by Clough and Wu (2022) to outline three clarifications to our theory of data network effects, concerning (1) conditions under which data network effects accrue, (2) the importance of theorizing shared data, and (3) the model’s ability to explain the cumulative effect of data-driven learning on value creation and value capture. Clough and Wu (2022)对我们关于人工智能（AI）和数据网络效应在创造用户价值中作用的文章（Gregory, Henfridsson, Kaganer, & Kyriakou, 2021）给出了一个有趣且发人深省的回应。我们欢迎围绕数据网络效应作为一种新的网络效应类别展开的讨论。在本回应中，我们基于Clough和Wu（2022）提出的观点，对我们的数据网络效应理论进行三点澄清，涉及（1）数据网络效应产生的条件，（2）共享数据理论化的重要性，以及（3）该模型解释数据驱动学习对价值创造和价值捕获的累积效应的能力。
KEY CONDITIONS FOR DATA NETWORK EFFECTS
数据网络效应的关键条件
Clough and Wu’s (2022) response calls for further clarification about the conditions under which we can refer to “data network effects.” As recently explained by Cennamo (2020), building on the ideas presented by Hagiu and Wright (2020), there are two key conditions for data network effects. The first condition is that learning from one user should translate into a better product or experience for other users, not just that single user. In other words, as more users use the product, the product must improve the experience for all users. The second condition is that the product experience enhancement from learning should happen fast enough to affect the current value of the product. It should benefit its current users, not the next product generation’s users. In other words, the product improves over the consumption lifetime with more users adopting it. Clough和Wu（2022）的回应呼吁进一步澄清我们何时可以提及“数据网络效应”。正如Cennamo（2020）最近基于Hagiu和Wright（2020）提出的观点所解释的，数据网络效应存在两个关键条件。第一个条件是，从一名用户那里获得的学习应该转化为对其他用户（而非仅仅是该用户）更好的产品或体验。换句话说，随着更多用户使用产品，产品必须改善所有用户的体验。第二个条件是，从学习中获得的产品体验提升必须发生得足够快，以影响产品的当前价值。它应该惠及当前用户，而非下一代产品的用户。换句话说，随着更多用户采用该产品，产品在其使用周期内会不断改进。
Clough and Wu (2022) suggest that the term “data network effects” is misleading, assuming that the two conditions stated above are typically not met. This view, however, overlooks the role of AI that is central to the activation of data network effects. As stated in our original article (Gregory et al., 2021), the widespread use of AI on modern platforms makes a significant contribution to fulfilling the two conditions for data network effects stated above. First, a core characteristic of AI is learning from individual cases to identify and translate patterns into predictive models that feed into the iterative improvement of products and experiences for other users, not just one single user from which data are collected and analyzed. Second, another core characteristic of AI is the ability to efficiently scale data-driven learning and instantly release the resulting improvements to the product experience to affect the current value of the product for each user. While we explained these characteristics of AI in our original article (Gregory et al., 2021), we had not explicitly linked them to the key conditions for network effects, and we thank Clough and Wu (2022), Hagiu and Wright (2020), and Cennamo (2020) for triggering us to add this clarification to the debate. Clough和Wu（2022）认为“数据网络效应”这一术语具有误导性，前提是上述两个条件通常不成立。然而，这一观点忽视了人工智能在激活数据网络效应中核心的作用。正如我们在原文（Gregory等人，2021）中所述，现代平台上人工智能的广泛应用对满足上述数据网络效应的两个条件做出了重大贡献。首先，人工智能的一个核心特征是从个体案例中学习，识别模式并转化为预测模型，这些模型不仅用于从收集和分析数据的单个用户，还能推动产品和体验对其他用户的迭代改进。其次，人工智能的另一个核心特征是能够高效地扩展数据驱动的学习，并立即将由此产生的改进应用于产品体验，从而影响每个用户当前的产品价值。虽然我们在原文（Gregory等人，2021）中解释了人工智能的这些特征，但我们没有明确将它们与网络效应的关键条件联系起来。我们感谢Clough和Wu（2022）、Hagiu和Wright（2020）以及Cennamo（2020）促使我们在这场辩论中补充这一澄清。
DATA: FROM FIRM RESOURCE TO SHARED DATA
数据：从企业资源到共享数据
Another key argument presented by Clough and Wu (2022) as a foundation for their ideas about value creation and capture is that the accumulated data derived from the installed base of decentralized users exists internal to the boundaries of the firm within centralized data structures. This argument resonates well within the realm of the resource-based view of the firm in which resources under control of the firm, including intangible resources such as databases, serve as potential sources of competitive advantage (Hall, 1993). However, data should not be treated as any other type of resource or production factor such as capital, labor, or oil (Parra-Moyano, Schmedders, & Pentland, 2020). We surmise that Clough and Wu’s (2022) argument potentially breaks down when taking into consideration at least three emerging characteristics of data (omitted in Clough & Wu, 2022, as well as our own original article). To understand data network effects and their role for value creation and capture, we suggest it is critical to pay closer attention to the following characteristics of data. Clough和Wu（2022）提出的另一个关键论点，作为其关于价值创造和获取观点的基础，是：来自去中心化用户安装基数的累积数据，在集中式数据结构中存在于企业边界内部。这一论点与企业的资源基础观（resource-based view of the firm）高度契合，在该理论中，企业控制的资源（包括数据库等无形资源）可作为竞争优势的潜在来源（Hall，1993）。然而，数据不应被视为任何其他类型的资源或生产要素，如资本、劳动力或石油（Parra-Moyano、Schmedders和Pentland，2020）。我们推测，当考虑到数据至少三个新兴特征时，Clough和Wu（2022）的论点可能会不成立（这些特征在Clough和Wu（2022）以及我们自己的原始文章中均被省略）。为了理解数据网络效应及其在价值创造和获取中的作用，我们认为，密切关注数据的以下特征至关重要。
First, data are seldom solely strategic resources that exist internal to the boundaries of any individual firm. In a great deal of contexts, they are also a medium of signification and carrier of facts and meanings that serve as a basis for learning and discovery (increasingly, via machine learning or combinations ofhuman and machine intelligence; that is, “metahuman systems”; Lyytinen, Nickerson, & King, 2020) from which new insight and knowledge can emerge. As Alaimo, Kallinikos, and Aaltonen (2020) explained, viewing data as media of signification and representation, alongside as resources, helps in identifying core qualities of data that are closely associated with value creation processes, helpingus further understand the role of AI and data network effects for the creation of user value. Core qualities of data include editability (e.g., through aggregation, filtering, reordering, and expansion of data), portability (e.g., through the adoption and diffusion of common standards for structuring and sharing data), and re-contextualizability (e.g., through combination of ground truth data with local domain expertise and knowledge) (Alaimo et al., 2020). By focusing on these inherent qualities of data, we should expect the focus of firms competing on big data, AI, and data network effects to shift away from data control and toward data sharing (Wixom, Sebastian, & Gregory, 2020), supported by AI alignment to manage diverse stakeholder interests (Wixom, Someh, & Gregory, 2020). This leads us to our next point about data ownership versus access. 首先，数据很少仅仅是存在于任何单个企业边界内部的战略资源。在大量情境中，它们也是一种意义媒介和事实与意义的载体，是学习和发现（越来越多地通过机器学习或人机智能的结合，即“超人类系统”；Lyytinen, Nickerson, & King, 2020）的基础，从中可以产生新的见解和知识。正如Alaimo、Kallinikos和Aaltonen（2020）所解释的，将数据视为意义和表征的媒介，同时也作为资源，有助于识别与价值创造过程密切相关的数据核心品质，从而进一步帮助我们理解人工智能和数据网络效应在创造用户价值方面的作用。数据的核心品质包括可编辑性（例如，通过数据的聚合、过滤、重新排序和扩展）、可移植性（例如，通过采用和推广用于数据结构化和共享的通用标准）以及可重新情境化（例如，通过将真实数据与本地领域专业知识和知识相结合）（Alaimo等人，2020）。通过关注数据的这些内在品质，我们应该预期，在大数据、人工智能和数据网络效应方面竞争的企业的关注点将从数据控制转向数据共享（Wixom, Sebastian, & Gregory, 2020），这一转变得到了人工智能对齐以管理不同利益相关者利益的支持（Wixom, Someh, & Gregory, 2020）。这就引出了我们关于数据所有权与访问权的下一个要点。
Second, data do not have to be owned (yet accessed) to learn and improve through the use of AI. With the proliferation of data sharing agreements and widespread adoption and diffusion ofstandardized interfaces for data exchange, so-called “application programming interfaces,” firms are increasingly able to leverage the portability of data to access and create value with data by training unique machine learning models, without necessarily owning and controlling the training data. Once a machine learning model has been trained, it is able to function independently from the training data with which it was fed and developed. Growing availability of open data sets and emerging markets for data with built-in appropriation regimes and guardrails for quality control and data provenance also contribute to the shift away from a focus on exclusive ownership of data that creates significant challenges for privacy protection (Thomas, 其次，数据不必被拥有（但需被访问）即可通过人工智能的使用来学习和改进。随着数据共享协议的普及以及用于数据交换的标准化接口（即所谓的“应用程序编程接口”）的广泛采用和传播，企业越来越能够利用数据的可移植性，通过训练独特的机器学习模型来获取和创造数据价值，而不必拥有和控制训练数据。一旦机器学习模型被训练完成，它就能独立于其训练数据发挥作用。开放数据集的日益增多，以及带有内置所有权制度和质量控制及数据溯源保障机制的新兴数据市场，也促使人们从专注于数据的排他性所有权转向其他方向，而专注于排他性所有权会给隐私保护带来重大挑战（Thomas,
Leiponen, & Koutroumpis, 2020). In fact, the emergence of data exchanges promises to serve as “platforms that gather data from many different sources and that allow third parties to run algorithms on these data” (Parra-Moyano et al., 2020). 莱波宁和库特鲁皮斯，2020）。事实上，数据交换的出现有望成为“汇聚来自许多不同来源数据的平台，并允许第三方在这些数据上运行算法”（帕拉-莫亚诺等人，2020）。
Third, a significant overarching development that will predictably shift the focus further away from internally managing and controlling the data collected from the installed base of users (e.g., Facebook platform) is the ongoing transition from platform to token economy, and the associated shift from data monopolies to data sovereignty (Voshmgir, 2020). With few exceptions (e.g., Cennamo, Marchesi, & Meyer, 2020), the vast majority of existing studies about network effects in the last two decades have been carried out in platform-based settings. As a result, these studies have assumed a centralized, if only virtually, cloud and data storage model as inherent in the business models of platforms such as Facebook, Google, Amazon, and YouTube. The gradual transition from client-server Internet facilitating interactions to the decentralized Internet facilitating agreements and value exchange, enabled by blockchain networks, first started to manifest in the form of peer-to-peer money without banks. The use of blockchain technology, however, is not limited to storage and transfer of financial value, and examples such as Filecoin that seek to incentivize distributed and decentralized data storage highlight that tokenization (a special form of digitization that has been described as a method that converts rights to an asset into digital tokens that can be securely bought, sold, and traded on blockchains; Sazandrishvili, 2020), when applied more broadly to transform economies and markets beyond trying to revolutionize money, can potentially reduce further the data monopolies of the contemporary platform economy that Clough and Wu (2022) assume. 第三，一个重要的总体发展趋势将不可避免地使焦点进一步远离对从已安装用户群（例如Facebook平台）收集的数据进行内部管理和控制，这一趋势是平台经济向代币经济的持续过渡，以及随之而来的数据垄断向数据主权的转变（Voshmgir, 2020）。在过去二十年中，除少数例外（例如Cennamo, Marchesi, & Meyer, 2020），绝大多数关于网络效应的现有研究都是在基于平台的环境中开展的。因此，这些研究假设了一种集中式（即使只是虚拟的）云与数据存储模型，认为这是Facebook、谷歌、亚马逊和YouTube等平台商业模式的固有特征。由区块链网络推动的从促进交互的客户端 - 服务器互联网向促进协议和价值交换的去中心化互联网的逐步过渡，最初以无银行的点对点货币形式显现。然而，区块链技术的应用并不局限于金融价值的存储和转移，像Filecoin这样旨在激励分布式和去中心化数据存储的例子表明，代币化（一种特殊形式的数字化，被描述为将资产权利转化为可在区块链上安全买卖和交易的数字代币的方法；Sazandrishvili, 2020）如果更广泛地应用于变革经济和市场（而非仅试图革新货币），可能会进一步削弱Clough和Wu（2022）所假设的当代平台经济的数据垄断。
VALUE CREATION AND CAPTURE AS A DUALITY
价值创造与捕获作为一种二元性
A significant point made by Clough and Wu (2022) is that data-driven learning (i.e., platform AI capability) enhances the firm’s capacity not only to create but also to capture value and, accordingly, “firms (must) strategically decide whether more creation or capture” is needed. We concur that the focus on value capture is a crucial addition to the conversation about how firms compete in a data-rich world. However, rather than viewing data-driven value creation and capture in terms of a trade-off, we argue that, in this context, the strategic balancing act between favoring value creation or value capture needs to be considered holistically in terms of a duality. The proposed model of data network effects in our original article (Gregory et al., 2021) offers a solid foundation to develop such an inquiry. Clough和Wu（2022）提出的一个重要观点是，数据驱动的学习（即平台AI能力）不仅增强了企业创造价值的能力，还增强了其获取价值的能力，因此“企业（必须）战略性地决定需要更多的创造还是获取”。我们同意，对价值获取的关注是关于企业如何在数据丰富的世界中竞争这一讨论的重要补充。然而，与其将数据驱动的价值创造和获取视为一种权衡，我们认为，在这种情况下，对偏向价值创造还是价值获取的战略平衡需要从二元性的角度进行整体考虑。我们在原始文章（Gregory等人，2021）中提出的数据网络效应模型为开展此类研究提供了坚实的基础。
Clough and Wu (2022) provide several convincing examples wherein digital platforms choose to leverage AI capability to maximize value capture rather than value creation. The authors allude to the role of one of the moderating variable in the data network effects model—namely, data stewardship (i.e., data quantity and quality)—in enhancing the platform’s ability to “manipulate” its users through predatory price discrimination and/or malicious user experience. The example of Electronic Arts employing data about user behavior to make the unlocking of popular game characters unachievable is particularly telling. Clough和Wu（2022）提供了几个有说服力的例子，其中数字平台选择利用人工智能能力来最大化价值捕获而非价值创造。作者提到了数据网络效应模型中的一个调节变量——即数据管理（即数据数量和质量）——在增强平台通过掠夺性价格歧视和/或恶意用户体验“操纵”用户的能力方面的作用。电子艺界（Electronic Arts）利用用户行为数据使热门游戏角色无法解锁的例子尤其具有启发性。
Such behaviors on the part of the platform, as Clough and Wu (2022) aptly demonstrate, are often viewed with disdain by users who, at the extreme, may choose to leave the platform. To theorize these unintended outcomes, and to better understand the compound effect of data-driven learning on value creation and value capture, the other two moderating factors of the data-network effects model need to be brought into the fold. Incorporating “performance expectancy” (a part of user-friendly design) into our explanatory account, for instance, would suggest that making the unlocking of the game characters unachievable undermines the users’ belief that the desired outcomes are attainable, reduces their engagement, and, ultimately, may diminish the firm’s capacity to capture value. Similarly, considering the effect of “personal data use” (an element of platform legitimation) could help us predict whether maximizing value capture by manipulating psychological state of the users is a good long-term strategy. If deemed “the wrong thing to do” in the backdrop of prevalent moral beliefs, such strategy puts the firm at risk of losing access to vital resources provided by investors, regulators, and partners. 正如Clough和Wu（2022）恰当论证的那样，平台的这类行为往往受到用户的鄙夷，极端情况下用户可能会选择离开平台。为了对这些非预期结果进行理论化分析，并更好地理解数据驱动型学习对价值创造和价值捕获的复合影响，数据网络效应模型的另外两个调节因素需要被纳入考量。例如，将“绩效期望”（用户友好型设计的一部分）纳入我们的解释框架，会表明使游戏角色解锁不可实现会削弱用户对目标结果可实现性的信念，降低其参与度，并最终可能削弱企业的价值捕获能力。同样，考虑“个人数据使用”（平台合法性的一个要素）的影响，有助于我们预测通过操纵用户心理状态来最大化价值捕获是否是一项良好的长期策略。如果在普遍的道德观念背景下被视为“错误的行为”，这种策略会使企业面临失去投资者、监管机构和合作伙伴提供的关键资源的风险。
As this discussion highlights, firms are increasingly under pressure and have technologies at their disposal (e.g., open digital infrastructures, decentralized data governance, and interoperability standards) to reduce the natural tendency in platform capitalism to decide the tension of data-driven value creation versus capture in favor of the latter. In contrast to Clough and Wu (2022), our original article (Gregory et al., 2021) would suggest that firms embracing new digital technologies and societal demands do not necessarily have to strategically decide whether more creation or capture is needed—value creation and capture can potentially be combined into a duality that views the two elements as interdependent, rather than separate and opposed. Viewing value creation and capture in terms of a duality shifts the unit of analysis and managerial emphasis from individual actors (the firm serving and exploiting its external customers) toward value exchange within the network of relationships between actors (the firm embedding itselfinto the network and middle of peer-to-peer transactions), consistent with the rise and gradual diffusion of blockchain technology as one of the foundations for the gradual shift from centralized platforms to decentralized networks as the bedrock of building the new economy (Pentland, 2020). 正如这场讨论所强调的，企业正面临越来越大的压力，并且拥有可支配的技术（例如开放数字基础设施、去中心化数据治理和互操作性标准）来减少平台资本主义中数据驱动的价值创造与攫取之间的张力，转而倾向于后者。与Clough和Wu（2022）的观点不同，我们的原创文章（Gregory等人，2021）认为，拥抱新数字技术和社会需求的企业不一定需要战略性地决定是需要更多创造还是更多攫取——价值创造和攫取可以潜在地结合成一种二元性，将这两个要素视为相互依存，而非分离和对立。从二元性的角度看待价值创造和攫取，将分析单位和管理重点从个体参与者（即服务和剥削外部客户的企业）转向参与者之间关系网络中的价值交换（即企业嵌入网络并处于点对点交易的中间位置），这与区块链技术的兴起和逐步普及一致，区块链技术是从集中式平台逐步向去中心化网络转变的基础之一，而后者是构建新经济的基石（Pentland，2020）。
BALANCING DIVERSE STAKEHOLDER INTERESTS
平衡不同利益相关者的利益
As a concluding remark, we would like to comment on Clough and Wu’s (2022) observation that “it is common practice to design platforms in ways that capture value for the platform owner at the expense of the total value being created.” The model in our original article (Gregory et al., 2021) predicts that this practice is not sustainable and may not be that common for all too long a time, not least due to new regulations that are expected. We would like to remind readers of the need and growing awareness among managers to balance diverse stakeholder interests in managing data network effects, in consistency with the view of “stakeholder governance as a process of finding ways to resolve stakeholder conflicts” (Amis, Barney, Mahoney, & Wang, 2020: 501). While we do think that the unfolding debate about data network effects has yet to deliver solid answers to this central problem of management in today’s economy, the notion of platform legitimation was a central argument in our original article. Firms are predicted by the model in our original article (Gregory et al., 2021) to create (and capture) value only insofar as they make appropriate use of the oftentimes-personal data collected on each user. In other words, firms must ensure the moral desirability of the use of personal data, which includes considerations of security and privacy, and ensure explainability of predictions made by machine learning algorithms fed with this data. Without ensuring such appropriate use of personal data, our model predicts that value creation will likely not sustain itself, in turn also taking away the basis for value capture, putting the firm’s long-term success at jeopardy. 作为结语，我们想对Clough和Wu（2022）的观察进行评论，他们指出“设计平台以牺牲总创造价值为代价来获取平台所有者价值是一种常见做法”。我们原文中的模型（Gregory等人，2021）预测，这种做法不可持续，且不会长期普遍存在，这在很大程度上是由于预计将出台的新法规。我们想提醒读者，管理者需要并日益意识到在管理数据网络效应时平衡不同利益相关者的利益的必要性，这与“利益相关者治理是解决利益相关者冲突的过程”的观点一致（Amis, Barney, Mahoney, & Wang, 2020: 501）。虽然我们确实认为，关于数据网络效应的不断展开的辩论尚未为当今经济中这一核心管理问题提供坚实答案，但平台合法性的概念是我们原文中的核心论点。我们原文中的模型（Gregory等人，2021）预测，企业只有在适当使用从每位用户收集的往往是个人的数据时，才会创造（并捕获）价值。换句话说，企业必须确保个人数据使用的道德可取性，这包括考虑安全性和隐私，并确保使用这些数据的机器学习算法所做预测的可解释性。如果不确保对个人数据的适当使用，我们的模型预测，价值创造可能无法自我维持，进而也会剥夺价值捕获的基础，使企业的长期成功面临风险。
REFERENCES
参考文献
Alaimo, C., Kallinikos, J., & Aaltonen, A. 2020. Data and value. In S. Nambisan, K. Lyytinen, & Y. Yoo (Eds.), Handbook of digital innovation: 162178. Northampton, MA: Edward Elgar Publishing.
阿拉伊莫（Alaimo, C.）、卡利尼科斯（Kallinikos, J.）和奥尔托宁（Aaltonen, A.）。2020。数据与价值。载于S. 南比桑（S. Nambisan）、K. 莱蒂宁（K. Lyytinen）和Y. 柳（Y. Yoo）编，《数字创新手册》：162178。马萨诸塞州北安普敦：爱德华·埃尔加出版社。
Amis, J., Barney, J., Mahoney, J. T., & Wang, H. 2020. From the editors: Why we need a theory of stakeholder governance—and why this is a hard problem. Academy of Management Review, 45: 499503.
阿米斯，J.，巴尼，J.，马奥尼，J. T.，& 王，H. 2020. 编者按：为何我们需要利益相关者治理理论——以及为何这是个难题。《管理学会评论》，45: 499-503.
Cennamo, C. 2020, July 16. Value preserving platform regulation: Network effects, platform value, and regulatory remedies [Workshop]. Digital Markets Competition Forum, Copenhagen Business School, Denmark.
Cennamo, C. 2020年7月16日。价值保留平台监管：网络效应、平台价值与监管补救措施[研讨会]。丹麦哥本哈根商学院数字市场竞争论坛。
Cennamo, C., Marchesi, C., & Meyer, T. 2020. Two sides of the same coin? Decentralized versus proprietary blockchains and the performance of digital currencies. Academy of Management Discoveries, 6. doi: 10. 5465/amd.2019.0044
Cennamo, C., Marchesi, C., & Meyer, T. 2020. Two sides of the same coin? Decentralized versus proprietary blockchains and the performance of digital currencies. Academy of Management Discoveries, 6. doi: 10. 5465/amd.2019.0044
Clough, D. R., & Wu, A. 2022. Artificial intelligence, datadriven learning, and the decentralized structure of platform ecosystems. Academy of Management Review, 47: 184189.
Clough, D. R., & Wu, A. 2022. 人工智能、数据驱动学习与平台生态系统的去中心化结构。《管理学会评论》，47: 184189.
Gregory, R. W., Henfridsson, O., Kaganer, E., & Kyriakou, H. 2021. The role of artificial intelligence and data network effects for creating user value. Academy of Management Review, 46: 534551.
格雷戈里，R. W.，亨弗里德松，O.，卡加内尔，E.，& 基里亚库，H. 2021. 人工智能和数据网络效应在创造用户价值中的作用。《管理学会评论》，46: 534551.
Hagiu, A., & Wright, J. 2020. Data-enabled learning, network effects and competitive advantage (Working paper). Retrieved from https://ap5.fas. nus.edu.sg/fass/ecsjkdw/ data%20enabled%20learning%20june2020.pdf
哈吉乌，A.，& 赖特，J. 2020. 数据驱动的学习、网络效应与竞争优势（工作论文）。Retrieved from https://ap5.fas. nus.edu.sg/fass/ecsjkdw/ data%20enabled%20learning%20june2020.pdf
Hall, R. 1993. A framework linking intangible resources and capabiliites to sustainable competitive advantage. Strategic Management Journal, 14: 607618.
霍尔，R. 1993. 一个将无形资源和能力与可持续竞争优势联系起来的框架。《战略管理杂志》，14: 607-618。
Lyytinen, K., Nickerson, J. V., & King, J. L. 2021. Metahuman systems = humans $^ +$ machines that learn. Journal of Information Technology, 36: 427445.
Lyytinen, K., Nickerson, J. V., & King, J. L. 2021. 元人类系统 = 人类 + 会学习的机器。《信息技术杂志》，36: 427445。
Parra-Moyano, J., Schmedders, K., & Pentland, A. 2020. Shared data: Backbone of a new knowledge economy. (Working paper). Retrieved from https://wip.mitpress. mit.edu/pub/yvy3qigg/release/1
Parra-Moyano, J., Schmedders, K., & Pentland, A. 2020. 共享数据：新经济知识的核心支柱。（工作论文）。Retrieved from https://wip.mitpress. mit.edu/pub/yvy3qigg/release/1
Pentland, A. 2020. Building the new economy: What we need and how to get there. (Working paper). Retrieved from https://wip.mitpress.mit.edu/pub/ aps1hrbe/release/7
彭特兰（Pentland, A.）. 2020. 构建新经济：我们需要什么以及如何实现。（工作论文）。从 https://wip.mitpress.mit.edu/pub/aps1hrbe/release/7 获取。
Sazandrishvili, G. 2020. Asset tokenization in plain English. Journal of Corporate Accounting & Finance, 31: 6873.
Sazandrishvili, G. 2020. 用通俗英语讲解资产代币化。《公司会计与金融杂志》，31: 6873。
Thomas, L. D. W., Leiponen, A., & Koutroumpis, P. 2020. Markets for data. Industrial and Corporate Change, 29: 645660.
托马斯（Thomas, L. D. W.）、莱皮宁（Leiponen, A.）和库特鲁皮斯（Koutroumpis, P.）。2020年。数据市场。《产业与公司变革》（Industrial and Corporate Change），第29卷：645-660页。
Voshmgir, S. 2020. Token economy: How the Web3 reinvents the internet. Berlin, Germany: BlockchainHub Berlin.
Voshmgir, S. 2020. Token economy: How the Web3 reinvents the internet. Berlin, Germany: BlockchainHub Berlin.
Wixom, B. H., Sebastian, I. M., & Gregory, R. W. 2020. Data sharing 2.0: New data sharing, new value creation, research briefing. Boston, MA: MIT Sloan Center for Information Systems Research.
维科姆，B. H.，塞巴斯蒂安，I. M.，& 格雷戈里，R. W. 2020. 数据共享 2.0：新的数据共享，新的价值创造，研究简报。马萨诸塞州波士顿：麻省理工学院斯隆信息系统研究中心。
Wixom, B. H., Someh, I. A., & Gregory, R. W. 2020. AI alignment: A new management paradigm, research briefing. Boston, MA: MIT Sloan Center for Information Systems Research. 维科姆，B. H.，索梅，I. A.，& 格雷戈里，R. W. 2020. 人工智能对齐：一种新的管理范式，研究简报。马萨诸塞州波士顿：麻省理工学院斯隆信息系统研究中心。
Robert Wayne Gregory Ola Henfridsson University of Miami 罗伯特·韦恩·格雷戈里 奥拉·亨弗里德松 迈阿密大学
Evgeny Kaganer Moscow School of Management SKOLKOVO 叶夫根尼·卡加涅尔 莫斯科斯科尔科沃管理学院
Harris Kyriakou ESSEC Business School https://doi.org/AMR_20210111 Harris Kyriakou ESSEC Business School https://doi.org/AMR_20210111
Copyright of Academy of Management Review is the property of Academy of Management and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder’s express written permission. However, users may print, download, or email articles for individual use. 《管理学会评论》的版权归管理学会所有，未经版权所有者明确书面许可，其内容不得复制、发送至多个网站或发布到邮件列表。不过，用户可打印、下载或通过电子邮件发送文章用于个人用途。