Impressions of an algorithmic new life inevitably have their boundaries. In a recent analysis of ‘eight public engagement experiments’, Annette Markham (2021) explored in some detail the limits that solidify within framings of algorithmic futures. Markham found that there were strong and persistent frames in place that secured certain visions of future devices, data and algorithmic social formations. Yet these visions are not simply accepted without question. The discussions at these events, Markham found, revealed that, when prompted, a critical understanding of the platforms and technologies could be readily articulated by the participants. Despite the critical ruminations of these participants, Markham also found that the framing of particular future scenarios persisted and carried with them a strong sense of inevitability. The future was somehow fixed. There was something obdurate about these framings; something that made them seem unavoidable. Indeed, Markham’s (2021: 384) argument is that such ‘discursive patterns continually strengthen the dominant frames of inevitability and powerlessness’. This would suggest that algorithmic thinking, as we discussed in Chapter 1, has a strong sense of the future inscribed within it. More than this though, it is also suggestive of how robust and irresistible those framings have become.

Continuing with some of the themes covered so far, this chapter explores the way that algorithmic thinking comes to have limits or boundaries that constrain social forms. Or, more specifically, it reflects on how the limits of the known and the knowable are an active part of the tensions of algorithmic thinking. To be clear from the outset, this chapter is not trying to position or establish those limits, nor is it claiming that they are fixed and secure, rather it explores the types of tensions that arise at such boundaries and how those tensions might be understood. As Markham identified, however sturdy these limits might appear there remains scope for them to shift if provoked to do so, with the possibility to imagine alternatives outside those existing frames, especially if the means are found to support and encourage different ways of thinking about the possibilities. As I will focus upon in this chapter, these sites of movement and tension are associated with what is known and what is thought to be knowable.

Back in 1994, reflecting on the direction of rapidly advancing neural network technologies, the Nobel Prize-winning neural systems expert Leon N. Cooper mused on what the future might yet hold:

This juxtaposition of eras, as imagined a quarter of a century ago, hints at how the advancing computer science of the time, especially in its use of brain science as a foil, was beginning to see a future in which machines would hold escalating forms of intelligence. Where the past hundred years had been defined by machines, the coming hundred years, Cooper predicted, would be defined by how those machines were to become intelligent. With this, he imagined, would come an automated form of reasoning; an automation of the sensible. An advancing era of knowing was positioned on near the horizon, the new life of evolving machines was perceived to be just around the corner and it would bring with it automated forms of reasoning (see Chapter 1).

Cooper frames this in terms of what he imagined at the time would be a growing ability for computational reason. He also anticipated a relative and growing comfort with these forms of intelligence and what they might be used to achieve. His observation was that:

Without wanting to sound too epochal, it could be said that we are living in algorithmic times. We may not want to go so far, and I find myself trying to resist the temptation, but it has become hard to draw any other conclusion. The type of ‘programmed sociality’ to which Taina Bucher (2018: 4) has referred has become impossible to deny, especially as the algorithm ‘induces, augments, supports, and produces sociality’. Different types of algorithms have come to have very large-scale social consequences, ranging from shaping what people discover, experience and consume as they go about their mediated lives, through to financial decisions, trading choices, differential treatment around insurance, credit and lending and then on to housing, border controls and a wide array of other processes and decisions that just simply cannot be listed. Such features of the contemporary landscape have led to the compelling conclusion that this is a type of ‘algorithmic culture’ (Striphas, 2015) cultivated within the structures and relations of an ‘algorithmic society’ (Peeters and Schuilenburg, 2021). Given these circumstances we may even now be living an ‘algorithmic life’ (Amoore and Piotukh, 2016). Such conclusions are certainly merited. Part of the reason we might think of these as algorithmic times, if you would let me stick with that slightly hyperbolic phrasing for a moment longer, is just how long the list would be if we were to try to itemize every single way that algorithms have social or individual consequences. And even then, because of their often-invisible roles and the sheer complexity of overlapping systems, the list would be impossible to complete. The algorithm has become too enmeshed in the social world for it to be untangled from it.

Yet, far from being sleek and uncontested technologies informed by mutually recognized ideals or shared notions of progress, these algorithmic times are fraught with tensions. Algorithmic thinking is tense. This book is concerned with elaborating and conceptualizing some of these competing forces. More than this though, this book argues that these algorithmic times can only really be grasped if we are to engage with the specific nature of these strains. Understanding algorithmic times requires a sustained focus upon the tensions of algorithmic thinking.

Published in the 7 March 2020 edition of the Financial Times, the private banking firm Investec ran a full-page advert that seemingly sought to defend human decision-making. Going against the apparently relentless tide of automation, it gestures towards the pitfalls of an irrational and inflexible form of algorithmic thinking. The advert poses what appears to be a rhetorical question: ‘Who would be most likely to grant you a mortgage? An algorithm? Or a human being?’ In the unlikely case that the reader is unsure as to their position on this question, the background is filled by a monochrome photo of a comfortably seated human – it is not clear if they are the imagined customer or a representative of the lenders.

This particular advert is suggestive of two related things. First, by responding to it directly the advert highlights the established materiality of algorithmic social ordering. In order to function its key message requires there to have been some form of existing encounter with algorithmic structures of some sort. Second, it is indicative of the way in which the very notion of the algorithm has moved into public consciousness (as discussed in Beer, 2017). The apparent rush towards being algorithmic, in which organizations seek to present themselves as devolving powers to the apparent neutrality, objectivity and heightened efficiency of algorithms, creates opportunities for others to present themselves as providing an alternative. This is not an alternative to being algorithmic, I would add, it is more often simply a different version of it. In other words, the push towards algorithmic properties creates a space in which the human can be knowingly and actively reinserted into these systems. In the case of the Investec promotion, the attempt is to appear to be algorithmic while not abandoning a sense of human values. In other words, it is an attempt to actively seek to present this as an organization that uses automation without appearing to be too automated. There is an active avoidance of that particular boundary. Investec are, it would seem, aiming to avoid overstepping the perceived limits of algorithmic thinking.

In the summer of 2018 the technology company Hdac ran a television advert depicting their version of the ‘smart home’. With clean lines and neutral colours, the automated home space was a picture of hyper-functional minimalism. In many regards it was an entirely unremarkable advert – its style and tone were comparable with typical technology company promotions. Despite the familiar stylistic features, it was the advert’s very prominent mentions of ‘blockchain’ that was particularly notable. For the first portion of the advert a small message appeared at the bottom of the screen telling the viewer that ‘Hdac Technology is building the future with the blockchain solution’. This talk of future-building immediately returns us to Lefebvre’s impression of the new life as discussed in Chapter 1 – the blockchain home is also just around the corner, it would seem. Later in the advert the voiceover reiterates the central message, adding that the ‘Hdac Technology platform is smart and secure thanks to the blockchain solution’. This text appeared again at the bottom of the screen. It is clear that the main message of the advert is that blockchain is responsible for enabling the various visions of convenience and technological adaptability being depicted. The blockchain is also given responsibility for ensuring the security of these spaces. Despite its prominence in the messaging, the advert did not go on to say what blockchain was, nor did it mention its functionality or how it was to be applied. Blockchain instead seemed to stand in for a secure (while nonspecific) technical apparatus. Blockchain was itself the message. This advert is illustrative of how blockchain is associated with notions of ideal types of data security; it is also illustrative of how the term blockchain can even act as a byword for technical systems that ensure this security.

Given its unexplained insertion in this advert and the fact that it is a term that can be used without the need for definition, it would seem that the concept of blockchain has already become a fixture of a wider technological and perhaps even public and media discourse (as discussed in Chow-White et al, 2020). It is a recognizable term. It is an established signifier.

Characterized by a transparent regulatory system and a laissez-faire type of openness to commerce, Hong Kong is a major international business centre with over 4,000 regional headquarters and offices of leading multi-national corporations that attract significant international data flows (Hong Kong Government, 2021). A technically sophisticated smart city with high levels of internet connectivity (e.g. mobile subscriber penetration exceeds 280 per cent, and the average internet downloading speed ranks among the top three fastest cities in the world) and an extensive external telecommunications infrastructure comprising 11 submarine optical fibre cable systems (providing about 90 TB/s capacity), 20 overland optical fibre cable systems and 11 satellite systems. Hong Kong also boasts of being one of the safest cities for data centres in Asia (Hong Kong Innovation and Technology Bureau, 2020).

Hong Kong's approach to managing cross-border data flows is in line with its approach to commerce – and is based on a ‘by default’ policy of patchwork legislation and FTA commitments. The question is whether Hong Kong can remain thriving with a by default approach or whether it needs a coordinated policy. To answer the question, we must first assess Hong Kong's objectives and its uniqueness.

This chapter examines Hong Kong's ambitions to be a leading business centre, the importance of data to achieving these goals and salient data-flowrelated issues, whether Hong Kong should adopt a coordinated data flow policy and the key points Hong Kong must consider in deciding whether or not to formulate such a policy.

The Ambitions of Hong Kong

Hong Kong is a successful financial gateway to China and a regional, if not global, finance hub, featuring a high degree of liquidity. Home to many financial institutions providing a wide range of financial products and services to local and international investors, Hong Kong is widely viewed as a leading financial centre in Asia. Its legal system is an important ingredient in its success in financial services, with investors and businesses attracted by the security and predictability of the territory's strong adherence to the rule of law. Illustrating these points was Hong Kong's rating in the 2020 Heritage Foundation's Index of Economic Freedom, where the territory ranked 2nd among 186 economies in terms of economic freedom and 15th in the category ‘Government Integrity’, under the heading ‘Rule of Law’ (Heritage Foundation, 2021).

Data is now one of, if not the world's most valuable resource. The dramatic adoption of data-driven applications across economic sectors has made data and the flow of data so pervasive that it has become integral to everything we as members of society do – from conducting our finances to operating businesses to powering the apps we use every day.

Flows of knowledge and technology are at the centre of new networks driving production, innovation and opportunities. Data and its flow across borders are the lifeblood of the internet economy and significantly impacted various industrial sectors and the global economy. The growth of networks has allowed businesses to change how they structure and manage their design, production, marketing, customer support and other processes in order to optimize competitiveness and innovation.

Data is also fundamental to machine learning (ML)-based products and services. While developments such as the internet of things (IoT) and data-driven artificial intelligence (AI) systems will hasten the growing importance of the digital economy, these innovations are both strategic and sensitive due to potential externalities for businesses, governments, non-profits and more generally for society. These externalities generate new forces within industries that alter competitive dynamics, and consequently, result in new strategies and management practices. These strategic considerations have substantial implications for policymakers and regulators (Valavi et al., 2021).

Although data plays an outsized role in the evolution of a country's business and living environment, the breadth and impact of data make it difficult for a nation to ‘regulate’ per se. This is the result of numerous factors, including the growing number of interconnected devices that individually can generate ever-larger amounts of data with each new product line. Generation and application of data are now occurring in ways that had been unimaginable and/or impractical just a few years earlier. Domestic legislatures are not only struggling to catch up with the complexities associated with the technology but also in formulating standards and rules which are effectively targeted and do not result in serious unforeseen consequences. This has proven to be challenging, and trade-offs between, say, privacy and industrial policy or intellectual property (IP) protection and the development of AI must be made.

In August 2021, the Financial Secretary Paul Chan stated that Hong Kong should have a strategic plan for formulating new development directions (Xinhua, 2021). Developing a data governance framework for cross-border data flows that balances competing interests, enables efficient operations, does not stifle business while keeping anti-competitive behaviour in check and is flexible enough to handle future innovations and challenges would be an excellent place to start. But to do so requires awareness of the degree to which systems are interconnected and the implications (and fragilities) of the resulting complex interdependencies and conflicts.

Spotting the predicted effects of data interconnections can sometimes be done with relative ease – for instance, concluding that combining personal data and mining it for behavioural insights and other information could clash with privacy protections is not hard (European Commission, 2021c). On other occasions, the link or the breadth of the link may not be so obvious. This will be the case with Europe's proposed but yet to be ratified Artificial Intelligence Act (AIA), which will place any product that contains AI into prohibited, high-risk and limited categories, prescribing obligations for each category. The AIA will also ban the use of AI for subliminal psychological manipulation as well as prohibit public authorities from using real-time biometric surveillance and building AI-powered social scoring systems. The interconnection, while less obvious, is no less significant with one report suggesting that the AIA could cost the EU €31 billion over the next five years in added compliance and other expenses, deter investment into European AI start-ups, slow the digitization of the economy and encourage a brain drain of European entrepreneurs,data scientists and AI developers to locations with fewer bureaucratic hurdles (Mueller, 202b).

But there was another interconnection the review did not include, that is, how the EU's underlying policy of cooperation with ‘like-minded countries’ who espouse European rules and values (particularly concerning the protection of privacy and personal data) could disadvantage European firms (European Commission, 2020d). These could occur in numerous ways. For instance, there is a real risk that the new restrictions will stif le data f lows from countries the EU does not deem to be satisfactorily ‘like-minded’.

Formulating a policy on cross-border data flows is challenging for several interconnected reasons. The increasing number of devices connecting on the IoT is creating a staggering amount of data, and with every successive product iteration, each device can generate more data. Increased levels of connections breed innovation – for example, data sent across borders by these devices can be mined for valuable customer insights that can, in turn, not only lead to new services or products but could also be combined with other data in innovative applications that had not even been conceived just a few years earlier. And, of course, cross-border communications enable interactions that might result in yet more innovation.

Global interconnections can also breed complexities, and legal, ethical, societal and cultural issues come into play when data crosses borders and standards are not harmonized. Some countries, such as the United States, advocate almost total free data movement with regulation only to the extent necessary. Others, such as China, adopt a model of central control, and companies wanting to do business in China must adhere to their data rules, whether this means restricting certain data from leaving the borders, locating infrastructure within the country or even dictating corporate behaviour (such as prompting companies to suspend plans to list on stock exchanges for fear of infringing ever-changing rules of what data can or cannot leave a country ’s borders) (Gao, 2018; Lim, 2021).

Before a jurisdiction can successfully formulate a policy on data and data flows, it needs to be able to navigate the landmine of converging and crisscrossing objectives. To do so, officials must understand the nature of modern data, what makes data valuable and what makes its management so tricky: that a solid technological infrastructure is important is obvious, that subtle changes in the law may produce unexpected results may not be obvious.

Before delving into the different models of data governance (in Chapter 3) and evaluating frameworks for managing cross-border data (in Chapters 4 and 5), we first need to explain what makes data different from other types of assets and why the sharing of data is so important in the modern global economy.

The internet is global and borderless, but regulators in most jurisdictions attempt to ensure the relatively free flow of data and the maintenance of ‘trust’ in the system. Regulators mostly act independently and without any direction from the international community, resulting, unsurprisingly, in a variety of incongruent and incompatible regulations.

The internet relies on a global flow of data, with a growing number of dataintense digital services, utilizing a high amount of data in their production processes. In many cases, the data cross borders multiple times before the service is consumed (van der Marel and Ferracane, 2021), and regulators must therefore appreciate the necessity of cross-border data flows from a trade perspective (World Economic Forum, 2019). The free flow of data is critical to internet-based services (e.g., cloud computing) and e-commerce and essential to the development of technologies that rely on access to high-quality data that often resides in more than one territory. Cross-border data transfers are also important for less obvious reasons, including cross-border health, investigations and medical emergencies (Moorthy, 2020). An example of the last point is the 2013–2016 Ebola virus outbreak in West Africa where deficiencies in data-sharing mechanisms brought the question of data access to the forefront of the global health agenda (Giles-Vernick, et al., 2016).

Through the development and deployment of these data-reliant technologies and solutions, nations and the companies working within those borders can expect to derive increased economic and social values from crossborder data flows. In fact, the data economy has risen even more rapidly in a COVID-19-disrupted world in its importance for new economic growth opportunities (Casalini et al., 2021).

Both countries and industries recognize the importance of keeping global data flows as unrestricted as possible. As this chapter discusses, most governments are increasingly keen to ensure that their rules of data governance are consistent with those of other countries and have begun negotiating trade agreements to ensure the compatibility of standards. Likewise, companies seeking influence intensely lobby at the domestic and international levels for relatively unrestricted data flows. In establishing policies for cross-border data transfers, governments must also be aware that they can only attract inbound transfers of data and information technologies if people, businesses and other governments trust their system and regulatory framework.