Platforms and Artificial Intelligence

Over the past 5 years, platforms have emerged as a key organizational concept, notably due to the ubiquity of digital. This chapter aims to: (1) analyse the key economic challenges related to the platform phenomenon and (2) review the current analytical approaches before (3) proposing an integrated approach to how platforms work in practice. The chapter then concludes with an exploration of the business and policy implications for innovation.

The rapid growth of China’s digital economy features the rise of Internet-based digital platforms, which provide a vast sphere where massive business value is created. As Internet-based digital platforms have developed into the dominant organizational form of the digital economy, it is worthwhile to review the evolution of such platforms, reflect on their structural characteristics, and explore possible future directions. In this chapter, we examine the four stages of the evolution of Internet-based digital platforms in China, through which the initial four basic functional types of digital platforms have been continuously fusing and new forms of platforms have been emerging and shaping up. We then discuss the current situation of digital platforms in China, using a theoretical framework developed from an institutional perspective, based on which future trends and possible research opportunities are identified.

Due to the central role played by platforms in the development of new technologies such as AI, the study concentrates on patents applicants that collaborate with platforms, considering that collaborations create an essential link for the transmission of knowledge. Using patent data and a panel of 22 platforms and 207 worldwide applicants that share the characteristics of having collaborated at least once with platforms in their life, indicators of knowledge creation, knowledge stock and knowledge spillovers are constructed. Distinguishing applicants by organisation type (platforms, large firms, SMEs and universities), the effect of different types of spillovers on each of the different categories’ knowledge is studied via negative binomial regressions. While we found evidence in favour of knowledge diffusion between firms in the whole sample, not every category benefits from these spillovers: large firms are found to be those that benefit more and from a greater number of spillovers, followed by SMEs, while universities receive very little benefits from spillovers in terms of knowledge creation, and more importantly platforms, despite being the main creators of knowledge in the AI sector, do not benefit of spillovers, which attests to their internalised knowledge strategy with regard to AI investment.

Artificial intelligence (AI) comprises a range of methods that allow to extract and exploit more information from larger amounts of data. AI is recognised as a transformative technology for most economic activities and for government operations. It might be a driver of economic growth in the coming decades, contribute to improved well-being and healthcare, but also trigger churning on the labour market and possibly widening income inequalities. Hence the interest of all governments in this technology and the design and deployment of significant policy agendas. National policies aim at accelerating the pace of development of AI and its deployment for serving the public good, while mitigating its adverse effects. This chapter reviews the economic characteristics of AI and the policy agendas; it examines 12 selected national initiatives in the field of AI.

By 2023, it is estimated that more than 29 billion devices will be connected to Internet Protocol networks across the globe, most of which will be creating data in real time. Once combined with big data, 5G or artificial intelligence, they enable the automation of entire business processes, paving the way to a data-driven economy. The study analyses patent applications that constitute the building blocks of this Fourth Industrial Revolution (4IR). The data presented offer insights into which countries, companies and regional clusters are leading the way in 4IR technologies. By highlighting the fields that are gathering momentum and the cross-fertilisation taking place between these fields, it provides a guide for policy and business decision makers to direct resources towards value creation in the digital era.

The acquisition of artificial intelligence (AI) systems is a relatively new challenge for the international community, but one organization that has placed a major interest in acquiring AI is the United States Department (U.S.) of Defense (DoD). This book chapter will focus on the DoD and its challenges in the development and fielding of major AI systems to glean lessons from addressing these challenges that might benefit the international community of project managers who must manage AI acquisition programs. The chapter will focus on the standard DoD acquisition program management methodology, i.e., Earned Value Management (EVM), and how it might be improved through incorporation of two methodologies, i.e., Integrated Risk Management (IRM) and Knowledge Value Added (KVA), in the managing of complex DoD information technology (i.e., AI) programs. This research compared and contrasted these three methodologies with the goal of demonstrating when and how each method can be applied to improve the acquisitions lifecycle for AI systems. Finally, the results of this study can also be applied to for-profit and other non-profit organizations throughout the international community.

Artificial intelligence is sought to have transformative power of modern economies. This chapter discusses various aspects of this transformation, at the job, firm and macro-economic level and analyses how these changes affect the extent to which AI-based tools can trigger faster growth in productivity. It analyses why productivity has not accelerated yet, despite the exponential growth in the field of AI over the past decade. The paper makes a distinction between innovation and adoption and shows how rising inequality is both the result of a shift to AI technologies and a reason for slow adoption. The paper also provides an overview of various policy options to address the issues identified in this paper, ranging from social protection to competition policies that are necessary to address both frictions related to job reallocation and rising inequality.

Artificial Intelligence (AI) is not the solution to everything. It is important to understand that computational models in general and AI in particular face some limitations. Overcoming these limitations requires a lot of human expertise. As a consequence, AI systems are largely made of human intelligence. This chapter gives an overview of how AI programs work, what they can do and what are their limitations. It discusses the relation of such platforms to human skills.

This chapter addresses to issue of designing the next generation of comptences in relationship to AI development. Taking a foresight perspective, the chapter distinguished three level for such a design: (1) The macro level: this relates to policies to be developed at governmental or inter-governmental level. Problems include those related to employment, privacy, competition (especially with regard to the dominance of systemic platforms) and institutions (democracy); (2) The global (platform) level: this is important because these organizations will play a key role in designing value spaces and their related competences; (3) The territorial (regions, cities, communities) level, since value will be organized around territories, in more independent ways; and the individual firm level (excluding global platforms), since firms are the unit of decision making, and structure the organization of resources in the market economy.

The novel coronavirus infection (COVID-19) has had a profound impact on corporate management. The introduction of telework to curb the spread of infection continues to have a profound impact on the form of collaboration within companies. However, even after the first declaration of the state of emergency was lifted, some companies continued to maintain and develop telework, while others returned to traditional office work. In this paper, based on an “Emergency Organization Survey” conducted in April 2020 on the changes in work styles in these companies, we will clarify what challenges Japanese companies are facing in introducing telework. We will also examine the role that corporate digitization, also known as digital transformation (DX), has played in the introduction of telework and changes in business processes.

Real-time management of Artificial Intelligence (AI) becomes a central enabling function for coping with the rapid market changes and increasing demands of stakeholders. But diverse sensing of real time makes it tricky for enterprises to adjust business processes towards a real-time-based era and build the temporal conditions needed for deploying AI in a humanistic manner. This chapter, therefore, introduces the concept of Fast and Flow in an AI engagement context. Fast and Flow encompasses two ideas: one considers time as a monetary asset that helps to increase value; the second does not seek to control time and does not define it on the clock scale, rather, it describes the sense of presence. By introducing the components of Fast and Flow interaction, we provide a cognitive psychology dimension to the management of AI and explore the balancing of Fast and Flow in three possible AI scenarios. The first scenario is “business as usual,” but faster and more complex; the second scenario is more focused on consumers and is based on an ideal combination of AI and Fast and Flow management; in the third scenario, there is an overflow of technology—AI is too fast, and people are unable to control it. Finally, we are asking what Fast and Flow management can do for a humanistic deployment of AI in enterprises and societies.

Increasing digitalization has created the concept of the digital divide—concerns about how no or inadequate internet access and use of related information could exacerbate existing socioeconomic gaps between industrialized and developing countries, metropolitan and rural areas, and more and less privileged individuals and groups. The revolution and disruption that artificial intelligence (AI) is expected to bring across all industries and fields of life are currently shifting these discussions toward access to and leveraging of AI, often referred to as the “AI divide,” involving competitive advantage, skillsets, development level and economic growth gap between countries, companies, universities, and individuals (There are discussions about more than one “AI divides”–see for example the three AI divides described in this WEF article referring to companies, skills and countries: https://​www.​weforum.​org/​agenda/​2018/​09/​the-promise-and-pitfalls-of-ai).

Related concerns about the current situation or the future are difficult to substantiate. Similarly, the known principle of data science “correlation does not mean causation” applies to this multifactorial issue. This chapter assesses the current AI divide based on scientific and patent publications related to AI as indicators of research and innovation output in the field. By considering the profile of the innovators and researchers, their affiliation, and geographies, it explores how different profiles and geographies already have access to necessary resources, showcase skills in the field of AI, and can or could deploy related applications. This chapter further explores existing policies and initiatives for building AI talent, strengthening AI-relevant skillsets and competencies, funding and strengthening AI research, offering incentives for establishing or attracting AI companies or further policies and measures to create an enabling environment, and leveraging the AI potential.

Patenting activity shows that AI-related research and innovation is rather concentrated both in terms of geographies as well as innovators. The United States and China are leaders in the AI innovation run—as origins of innovation and as locations of patent protection, and therefore as existing or potential markets. Background research related to policies in these jurisdictions and consultation with AI subject matter experts showed that this is largely due to the strength of their policy, education funding, and business ecosystem. Europe is ranked third, with the rest of Asia, Latin America, and Africa lagging behind. The scientific literature shows nevertheless that some research has been carried out in all these regions, but this may not be reflected—or at least not to its full extent—in related patenting activity, which tends to be more of an indicator of commercialization potential and related investment and industrial application. For this reason, it is important to look at patent and scientific literature data side by side before drawing any conclusions, as some countries’ strength in AI research is only or mainly reflected in the volume of scientific publications.

Looking at innovator profiles, there is a small number of ICT companies—mainly from the USA, China, Japan and the Republic of Korea, which lead patenting activity across all possible application fields. The area of transportation is an exception, with automotive industry representatives leading related activity. Moreover, these bigger AI players focus their patent filing strategy on a limited number of patent jurisdictions, indicating that the existing or potential AI markets for bigger AI players is from a commercial perspective and understanding limited to a rather small number of countries. Smaller entities tend to have very small patent profiles and be focused on their local markets.

The findings of the patent and scientific publications research show that a certain AI divide does exist if we consider it from an access and use perspective, looking at both geographies and profiles of AI researchers and innovators. Nevertheless, as AI is an emerging trend that several players are now joining, even the smaller activity across different countries and from different players indicates the potential which seems to already be there, and which can be enhanced and contribute to economic growth and development for all. Moreover, as AI is often based on open-source software and tools, access seems to be more democratized than other digital assets and tools, a factor that can even contribute to lessening the digital divide. A less obvious point for accessing and leveraging AI is the access to and ownership of training data which can facilitate or impede the development and applications of AI, making related policies important for how the future will look like for increasing or decreasing the digital divide.

The question for the future relates to the priorities of individual countries and regions do they envision themselves as competitors to the USA and China or do they prefer to identify their competitive advantage and focus on their own needs and strengths (See related considerations of Andrew Ng in Landing AI, AI Transformation Playbook (2018), available at https://​landing.​ai/​wp-content/​uploads/​2020/​05/​LandingAI_​Transformation_​Playbook_​11-19.​pdf; and in the WIPO Magazine interview (June 2019 issue) available at www.​wipo.​int/​wipo_​magazine/​en/​2019/​03/​article_​0001.​html); and what are other principles that need to be taken into consideration to ensure equitable AI? Different initiatives may enable AI talent building related skills, leveraging the AI potential with applications in industries of interest for the economies in question, in niche areas and adapted to local conditions. To fully realize the AI potential in terms of deployment, scaling-up, and applications, there is a need for sufficient AI which must be aware of the possibilities and limitations of AI and be led by evidence and real needs rather than hype. It is important to create the right conditions so that—similar to the aims of the UN Sustainable Development Goals—“no one is left behind” (Excerpt from Committee for Development Policy, See Official Records of the Economic and Social Council, 2018, SupplementNo.13(E/2018/33); https://​sustainabledevel​opment.​un.​org/​content/​documents/​2754713_​July_​PM_​2.​_​Leaving_​no_​one_​behind_​Summary_​from_​UN_​Committee_​for_​Development_​Policy.​pdf) in the AI revolution.

This study deals with the nature of a post-truth society, which is imminent, considering the widespread use of artificial intelligence (AI)-based information systems that use machine learning methods such as deep learning, as well as the social attitudes and responsibilities of organisations that develop, implement, operate and/or use those systems. In such a society, the truths about individuals, groups, organisations, communities, societies, nations, things, events and the world become meaningless or worthless; individuals are treated as black boxes to be manipulated and exploited by malicious AI-based system operators; and the four factors that erode accountability in computing—many hands, bugs, the computer as a scapegoat, and ownership without liability [Nissenbaum (Science and Engineering Ethics 2(1):25–42, 1996)]—worsen because of the unpredictability and uncontrollability of AI-based system behaviours, leading to a lack of responsibility and accountability in AI computing. To prevent the full emergence of a post-truth society or mitigate risks associated with such a society, and to restore responsibility and accountability to computing, organisations that are key players in AI computing must be required to proactively address ethical and social issues caused by the development and use of AI-based systems.

We present the motivation, design, outline, and lessons learned from an online course in scientific integrity, research ethics, and information ethics provided to over 2000 doctoral and engineering students in STEM fields, first at the University Paris-Saclay, and now expanded to an online MOOC available to students across the world, in English. Unlike a course in scientific domains, meant to provide students with methods, tools, and concepts they can apply in their future career, the goal of such a training is not so much to equip them, but to make them aware of the impact of their work on society, care about the responsibilities that befall on them, and make them realize not all share the same opinions on how should technology imprint society. While we provide conceptual tools, this is more to sustain interest and engage students. We want them to debate on concrete ethical issues and realize the difficulty of reconciling positions on contemporary dilemma such as dematerialized intellectual property, freedom of expression online and its counterparts, the protection of our digital selves, the management of algorithmic decision, the control of autonomous systems, and the resolution of the digital divide. As a bold shortcut, our course is about introducing and motivating Hegelian dialectics in STEM curricula, usually more bent on an Aristotelian perspective.