This last chapter summarizes most of the material in this book in a range of concluding statements. It provides a summary of the lessons learned. These lessons can be viewed as guidelines for research practice.

This chapter is dedicated to the memory of Sue Atkins, the Grande Dame of lexicography, who passed away in 2021. In a prologue we argue that she must be seen on a par with other visionaries and their visions, such as Paul Dirac in mathematics or Beethoven in music. We review the last half century through the eyes of Sue Atkins. In the process, insights of other luminaries come into the picture, including those of Patrick Hanks, Michael Rundell, Adam Kilgarriff, John Sinclair, and Charles Fillmore. This material serves as background to start thinking out of the box about the future of dictionaries. About fifty oppositions are presented, in which the past is contrasted with the future, divided into five subsections: the dictionary-making process, supporting tools and concepts, the appearance of the dictionary, facts about the dictionary, and the image of the dictionary. Moving from the future of dictionaries to the future of lexicographers, the argument is made that dictionary makers need to join forces with the Big Data companies, a move that, by its nature, brings us to the US and thus Americans, including Gregory Grefenstette, Erin McKean, Laurence Urdang, and Sidney I. Landau. In an epilogue, the presentation’s methodology is defined as being “a fact-based extrapolation of the future” and includes good advice from Steve Jobs.

The Conclusion chapter reiterates the book’s approach, focus and main points. It reminds the reader that the book has concentrated on local, provincial, peripatetic and otherwise relatively marginal sites of scientific activity and shown how a wide variety of spaces were constituted and reconfigured as meteorological observatories. The conclusion reiterates the point that nineteenth-century meteorological observatories, and indeed the very idea of observatory meteorology, were under constant scrutiny. The conclusion interrogates four crucial conditions of these observatory experiments: the significance of geographical particularity in justifications of observatory operations; the sustainability of coordinated observatory networks at a distance; the ability to manage, manipulate and interpret large datasets; and the potential public value of meteorology as it was prosecuted in observatory settings. Finally, the chapter considers the use of historic weather data in recent attempts by climate scientists to reconstruct past climates and extreme weather events.

Public health data available for research are booming with the expansion of Big Data. This reshapes the data sources for DOHaD enquiries while offering ample opportunities to advance epidemiological modelling within the DOHaD framework. However, Big Data also raises a plethora of methodological challenges related to accurately characterising population health trajectories and biological mechanisms, within heterogeneous and dynamic sociodemographic contexts, and a fast-moving technological landscape. In this chapter, we explore the methodological challenges of research into the causal mechanisms of the transgenerational transfer of disease risks that characterise the DOHaD research landscape and consider these challenges in the light of novel technologies within artificial intelligence (AI) and Big Data. Such technologies could push further the collating of multidimensional data, including electronic health records and tissue banks, to offer new insights. While such methodological and technological innovations may drive clearer and reproducible evidence within DOHaD research, as we argue, many challenges remain, including data quality, interpretability, generalisability, and ethics.

This chapter deals with how public policy can steer AI, by taking how it can impact the use of big data, one of the key inputs required for AI. Essentially, public policy can steer AI through putting conditions and limitations on data. But data itself can help improve public policy – also in the area of economic policymaking. Hence, this chapter touches on the future potential of economic policy improvements through AI. More specifically, we discuss under what conditions the availability of large data sets can support and enhance public policy effectiveness – including in the use of AI – along two main directions. We analyze how big data can help existing policy measures to improve their effectiveness and, second, we discuss how the availability of big data can suggest new, not yet implemented, policy solutions that can improve upon existing ones. The key message of this chapter is that the desirability of big data and AI to enhance policymaking depends on the goal of public authorities, and on aspects such as the cost of data collection and storage and the complexity and importance of the policy issue.

In this chapter, we describe the development of AI since World War II, noting various AI “winters” and tracing the current boom in AI back to around 2006/2007. We provide various metrics describing the nature of this AI boom. We then provide a summary and discussion of the salient research relevant to the economics of AI and outline some recent theoretical advances.

In this chapter, we take the production function enriched with AI abilities from Chapter 4, and apply it to study the implications for progress in AI on growth and inequality. The crucial finding we discuss in this chapter is that understanding the nature of AI as narrow ML and its effect on key macroeconomic outcomes depends on having appropriate assumptions in growth models. In particular, we discuss the appropriateness of assuming, as most standard endogenous growth models today do, that economies are supply driven. If they are not supply driven, then demand constraints, which can arise from the diffusion of AI, may restrict growth. Through this, we show why expectations that AI will may lead to “explosive” economic growth is unlikely to materialize. We show that by considering the nature of AI as specific (and not general) AI and making appropriate assumptions that reflect the digital AI economy better, economic outcomes may be characterized by slow growth, rising inequality, and rather full employment – conditions that rather well describe economies in the West.

With the emergence of big data and artificial intelligence, the feasibility of central planning has again become a popular topic. The essential feature of the planned economy is the use of systematic and institutional force to negate entrepreneurship and deprive individuals of the freedom to choose, especially the freedom to start a business and innovate. Can big data revive the planned economy? The essence of this question is: Can big data displace entrepreneurship? The impossibility of a big data-based planned economy is demonstrated from five perspectives (i.e. the nature of knowledge, the nature of entrepreneurial decisions, the distinction between risk and uncertainty, the importance of ideas, and the evolutionary view). In other words, big data cannot replace entrepreneurship. The false belief that central planning is possible with big data is extremely naïve.

The question of how to balance free data flows and national policy objectives, especially data privacy and security, is key to advancing the benefits of the digital economy. After establishing that new digital technologies have further integrated physical and digital activities, and thus, more and more of our social interactions are being sensed and datafied, Chapter 6 argues that innovative regulatory approaches are needed to respond to the impact of big data analytics on existing privacy and cybersecurity regimes. At the crossroads, where multistakeholderism meets multilateralism, the roles of the public and private sectors should be reconfigured for a datafied world. Looking to the future, rapid technological developments and market changes call for further public–private convergence in data governance, allowing both public authorities and private actors to jointly reshape the norms of cross-border data flows. Under such an umbrella, the appropriate role of multilateral, state-based norm-setting in Internet governance includes the oversight of the balance between the free flow of data and other legitimate public policies, as well as engagement in the coordination of international standards.