Is the managerial motivation behind artificial intelligence simply to reduce headcount by replacing employees with machines? If so, ‘’human-centric AI” is at best self-defeating nonsense. If MIT defines human-centric AI as “the design, development, and deployment of (information) systems that learn from and collaborate with humans in a deep, meaningful way” it is as important for management to learn where and how machine intelligence can enhance human potential. Human-centric AI depends upon corporate education’s ability to develop a clear understanding of what managers can bring to the table in increasingly digitally intermediated organizations.
What do business and engineering students need to know about AI in Management? In previous contributions to this series[i], we have explored themes ranging from “AI Readiness” and AI project management to Collaborative Intelligence and Digital Transformation. In this final installment, we will focus on three areas that we believe are the heart of any management curriculum on “human-centric” AI : the importance of Digital Ethics, the spurious link between AI and innovation, and the introduction of discussions on the value of AI.
Data ethics involves the study and adoption of data practices, algorithms, and applications that respect fundamental individual rights and societal values.[ii] If Ethics isn’t an attribute of either data or algorithms, ethical challenges do inherently arise wherever managers rely on data to take business decisions. Artificial Intelligence cannot be isolated from the larger economic and social challenges it has been designed to address when both the data and the algorithms reflect the visions, biases, and logic of human decision-making, Do these visions, biases and logic faithfully reflect the reality of business, and if not, how do these reflections color managerial views of customers, organizations and markets? Data-driven decision-making doesn’t imply choosing between objectively analyzing the data and subjectively influencing behavior but underlines the necessity to understand how data shapes human perceptions and initiatives.
In this light, the objective of discussions on data ethics isn’t in helping managers separate right from wrong, but in facilitating discussions over what constitutes acceptable digital practices. Five distinct challenges can be addressed and explored in management education. Because personal data becomes the fuel of the digital economy, the issues surrounding personally identifiable information, explicit consent, as well as the rights to access, to rectify, and to be forgotten merit attention. Algorithms have become the key to automating decision-making, but the extent to which managers should be held responsible for the resulting decisions remains open to debate. The questions of implicit bias are equally important, for individual and collective attitudes and preconceptions influence our use of data, cognition, logic and ethics. Technology’s impact on how and why we communicate should also be discussed — for our reliance on data has subtly modified our definitions of “trust”, “privacy”, “truth”, and “value”. Finally, should our commitment to data-driven decision-making elevate “scientism” over other forms of human intelligence?[iii]
HAI has 121 faculty members — over 80 percent are white, and almost as many are male
Stanford University’s Human-Centered Artificial Intelligence Institute (HAI) is an excellent example of both how difficult and how important examining ethical bias has become. Based on a position that “designers of AI must be broadly representative of humanity”, the University launched HAI this past year to “advance AI research, education, policy, and practice to improve the human condition.[iv]” Yet when the institute revealed the 120 faculty and tech leaders leading the initiative, all were educated in the world’s top business and engineering schools, over 80 percent were white, and almost as many were male. Does this group representative of population at large, or the different racial, cultural, and intellectual currents that have dominated the industry over the last five decades? Given the similarities in their backgrounds and education, to what extent are they capable of understanding the complexity of local ethical challenges? If the goal of digital transformation is to transform data into action, will the faculty be acting to reinforce the status quo, or to promote a more representative or eventually a more desirable diversity of beliefs of what the future may hold?
AI and Innovation
Is artificial intelligence proof of innovation, or at least a necessary condition for future innovation? Developing algorithms capable of imitating human behavior is one thing, building machines capable of innovating is another. Over the centuries, innovation has largely been context-dependent: different generations of innovation (Technology-push, market-pull, strategic integration, and networking) have been defined by the interplay between technological tools and market dynamics at specific points in time. Tidd and Bessant suggest that innovation today implies applying a product, process, position or paradigm to create value through solving economic or social problems.[v] Exploring how management can use machine intelligence to foster innovation constitutes a major opportunity for organizations to move the discussion of AI beyond data and algorithms and discuss how such technologies can be applied profitably to their business.
If artificial intelligence can reproduce human thought, corporate stakeholders should ponder why their managers need corporate education at all. Courses discussions can analyze the current state of machine intelligence, which at best is able to master repetitive tasks (Narrow AI), and the promises of “super artificial intelligence” capable of capable of reformulating the economic, social and political problems organizations are trying to solve. Seminars can explore why machine learning has been bound by the logic of solutions, whereas the roots of human innovation lie in uncovering the pervasiveness of apparent problems. Workshops can help managers understand how to leverage machine intelligence to identify discernable patterns in the data, and why they need to focus on the outliers that defy programmable logic. This apparent paradox may well condition management training in the foreseeable future: Although AI will never be proof of innovation, when used with proper training, AI constitutes an essential key in developing managerial curiosity, creativity and innovation.
When used with proper training, AI constitutes an essential key in developing managerial curiosity, creativity and innovation
The Chinese tech giant Alibaba is a case in point. In less than two decades Alibaba has produced ample evidence of its capacity to innovate: its platform-based business model was projected to generate $100 billion in revenue this year alone from its investments in retail, the internet, and technology.[vi] Alibaba’s management has redefined a “new retail” sector by leveraging Internet technologies and data intelligence to develop an open, technology-dependent, and value-centered ecosystem. [vii] Its vision of “smart business” is based on matching human and machine intelligence — each core business process is coordinated in an online network and use machine-learning technology to efficiently leverage data in real time. Management has leveraged Inventions like its cloud based Alink platform, applications like Alipay and Taobao, and its Ali assistant to create an innovative ecosystem of sellers, marketers, service providers, logistics companies, and manufacturers that outperform traditional business infrastructures. AI isn’t seen as an end in itself, but as a keystone in an innovation process of Datafying, Coordinating, Incentivizing, and Inspiring manufacturers and consumers to think differently about business.
AI and Value
What is the value of AI? “Value for money” is largely a pleonasm today, even if the issue of what values for money remains an open question. On one hand, consumers choose products and services using frames of reference that determine what they value. On the other, business processes today as designed as parts of a value chain putting into play the responsibilities of both producers and consumers. In both cases, perceptions of “financial value” are conditioned by human “values” such as quality, utility, fairness, proximity, and ownership. If the algorithms of AI can be easily trained to analyze and explore cost, they have to date have little ability to predict and/or influence human perceptions of value. Metrics capture what they were designed to measure, algorithms perform as they are taught, while people evaluate the worth of data based on personal experience.
Data has no intrinsic value, other than it how it is used to address social and economic challenges. Management training on AI needs to look beyond discussions of the pertinence of machine learning metrics like precision, accuracy, recall, and sensitivity to investigate the intricate relationships between financial value and human values. Case studies on the cognitive factors that influence purchasing decisions can help participants understand that decision-makers aren’t cold calculating machines acting on data but managers and consumers reacting to their feelings and insights. In introducing discussions of value into the study of artificial intelligence, corporate training can help management focus on how the interactions between machine and human agents influence productivity, effectiveness and quality. Machines neither create nor capture value, people do.
As long as humans purchase and consume, the essence of value will never be captured in an algorithm
Probing the managerial challenges of artificial intelligence provides a case in point. Limiting the study of AI to its technological components can blind management to the inevitable impact of AI on consumption, production, and investment. If value isn’t captured in the data itself, it is represented in the metadata whose intrinsic, extrinsic, and systemic dimensions define how data is organized and interpreted. The management of digital assets today requires a deep understanding of the semantics of closed taxonomies and open folksonomies inherent in valuating digital resources.[viii] Accounting for value rather than costs leads management to reflect on the link between AI and productivity, which in turn can lead to a redefinition of the production boundaries that determine the organization’s core activities in a digital economy. Finally, reformulating productivity metrics can feed discussions on incentives and profits and the opportunity to shift investment strategies from calculations of ROI for stockholders to the return on investments for all participants in the value chain.
What have we learned?
Corporate Training in Artificial Intelligence today needs to help management look beyond data and algorithms to explore how AI has transformed organizations and markets. The future success of business endeavors doesn’t depend on AI, but managers capable of leveraging digital technologies to enhance human potential. In this three-part series we have addressed nine themes that we believe should be at the core of any AI curriculum:
- AI and Strategy — how artificial intelligence influences the way we think about business.
- AI Readiness — the extent to which organizations are prepared to leverage the benefits of artificial intelligence
- AI Project Management — the management of AI experiments as projects aligned with an organization’s business objectives, context, and resources
- Collaborative Intelligence, — the design of networks of human and machine agents to foster the crowdsourcing of knowledge, expertise, and preferences
- Managerial Decision-Making — the focus on where and when machine learning can help managers transform data into action
- Digital Transformation — the vision that digital technologies don’t transform companies and markets, people do.
- Digital Ethics — the ethical challenges that arise wherever managers rely on data to take business decisions
- AI and Innovation — AI’s role in developing managerial curiosity, creativity and innovation.
- AI, Value and Values — the study of the valuation of AI
In each case, managerial skills rather than technical savvy prove fundamental in aligning the promise of AI with each organization’s strategy, resources, and culture.
Lee Schlenker is the Principal in the BAI http://baieurope.com and Professor of Business Analytics and Community Management. His LinkedIn profile can be viewed at www.linkedin.com/in/leeschlenker. You can follow the BAI on Twitter at https://twitter.com/DSign4Analytics
[iii] For example, emotional (interpersonal), linguistic (word smart), intrapersonal (self-knowledge) and spiritual (existential) intelligence
[iv] O’Neil, P.H. (2019), Stanford’s New Institute to Ensure AI Is ‘Representative of Humanity’ Mostly Staffed by White Guys, Gizmodo
[v] Tidd J., Bessant J. (2013) Managing innovation. Integrating technological, market and organizational change, Wiley, Chichester
[vi] These projections were forecast before the current economic recession.