Mutually Assured Education

Governments have begun to view AI compute infrastructures, including advanced AI chips, as a geostrategic resource. This is partly because “compute governance” is believed to be emerging as an important tool for governing AI systems. In this governance model, states that host AI compute capacity within their territorial jurisdictions are likely to be better placed to impose their rules on AI systems than states that do not. In this study, we provide the first attempt at mapping the global geography of public cloud GPU compute, one particularly important category of AI compute infrastructure. Using a census of hyperscale cloud providers’ cloud regions, we observe that the world is divided into “Compute North” countries that host AI compute relevant for AI development (ie. training), “Compute South” countries whose AI compute is more relevant for AI deployment (ie. running inferencing), and “Compute Desert” countries that host no public cloud AI compute at all. We generate potential explanations for the results using expert interviews, discuss the implications to AI governance and technology geopolitics, and consider possible future trajectories.

Brown, M. A., Gruen, A., Maldoff, G., Messing, S., Sanderson, Z., & Zimmer, M. (2024). Web Scraping for Research: Legal, Ethical, Institutional, and Scientific Considerations (arXiv:2410.23432). arXiv. https://doi.org/10.48550/arXiv.2410.23432

A nice overview of the different considerations, with recommendations, for researchers who may want to use web scraping.

As scraping for large-scale AI training becomes more common, scraping for other research purposes may become more challenging.

@mae

This report aims to go behind the headlines and beyond the simplified facial recognition (FR) debate that we see in so much of the media today. The report highlights the types of FR in use by policing and others around the world, and explores the big debates around that use. We interviewed 34 stakeholders – including policing practitioners, academics, regulators and suppliers – to share their views and expertise on this complex issue. This analysis includes an insightful foreword from National Police Chiefs' Council (NPCC) Chair Chief Constable Gavin Stephens.

Digital platforms have disrupted many sectors but have not yet visibly transformed highly regulated industries. This study of Big Tech entry in healthcare and education explores how platforms have begun to enter highly regulated industries systematically and effectively. It presents a four-stage process model of platform entry, which we term as “digital colonization.” This involves provision of data infrastructure services to regulated incumbents; data capture in the highly regulated industry; provision of data-driven insights; and design and commercialization of new products and services. The article clarifies platforms’ sources of competitive advantage in highly regulated industries and concludes with managerial and policy recommendations.

Policymakers face increased pressure to regulate digital markets to balance competition, privacy, and innovation. While traditional policy literature views regulation as a technical problem requiring specific interventions, there is significant debate about the appropriate solutions. This analysis synthesizes both technocratic and political perspectives, proposing a framework to predict effective regulatory remedies by examining common market structures.

Examining how U.S. regulation of media infrastructure over the past century—spanning broadband, digital platforms, and data centers—has eroded civil liberties and democratic principles, this book interweaves historical and interdisciplinary perspectives to elucidate how policies from an analog era continue to shape today’s digital governance. The work fosters a comprehensive understanding of the policy challenges faced and inspires potential restorative actions to realign infrastructure regulation with public interest and equity.

This talk analyzes how the US, China, and the EU shape the global digital order. The speaker examines different regulatory models: US market-driven, Chinese state-driven, and EU rights-driven approaches. The discussion highlighted shifts in tech regulation, including AI governance, geopolitical impacts, and the rise of digital sovereignty. Despite challenges, the talk emphasized the importance of democratic governance over critical tech policy decisions.

Abstract

Reinforced by the technological decoupling and the related battle for technological supremacy between the United States and China, telecommunication technology has become increasingly politicised. As the functioning of global telecommunication technology relies on interoperability and compatibility, government and private actors have strong incentives to shape the underlying standards in their economic and political interests. The International Telecommunication Union (ITU) and its Members play a central role in setting these standards for future telecommunication technology. Despite the ITU's importance in this field, relatively little is known about the organisation's work on standardisation and the actors behind it. This Policy Insight introduces a new data set on the involvement of over 800 government and private actors and their almost 50,000 contributions to ITU standardisation processes between 2000 and 2022. A descriptive analysis of the data set illustrates that particularly Chinese actors—Huawei, ZTE, China Mobile, China Unicom and China Telecom—have been actively driving the ITU's standardisation processes in the areas of transport, access and home but also future networks and cloud. The data set introduced here is envisaged as a source which allows researchers to study the reasons and implications for certain actors' involvement in the international standardisation of telecommunication and digitalisation.

This Article challenges the common view that more stringent regulation of the digital economy inevitably compromises innovation and undermines technological progress. This view, vigorously advocated by the tech industry, has shaped the public discourse in the United States, where the country’s thriving tech economy is often associated with a staunch commitment to free markets. U.S. lawmakers have also traditionally embraced this perspective, which explains their hesitancy to regulate the tech industry to date. The European Union has chosen another path, regulating the digital economy with stringent data privacy, antitrust, content moderation, and other digital regulations designed to shape the evolution of the tech economy toward European values around digital rights and fairness. According to the EU’s critics, this far-reaching tech regulation has come at the cost of innovation, explaining the EU’s inability to nurture tech companies and compete with the United States and China in the tech race. However, this Article argues that the association between digital regulation and technological progress is considerably more complex than what the public conversation, U.S. lawmakers, tech companies, and several scholars have suggested to date. For this reason, the existing technological gap between the United States and the EU should not be attributed to the laxity of American laws and the stringency of European digital regulation. Instead, this Article shows there are more foundational features of the American legal and technological ecosystem that have paved the way for U.S. tech companies’ rise to global prominence—features that the EU has not been able to replicate to date. By severing tech regulation from its allegedly adverse effect on innovation, this Article seeks to advance a more productive scholarly conversation on the costs and benefits of digital regulation. It also directs governments deliberating tech policy away from a false choice between regulation and innovation while drawing their attention to a broader set of legal and institutional reforms that are necessary for tech companies to innovate and for digital economies and societies to thrive.

Our current policy and research focus on artificial intelligence (AI) needs a paradigmatic shift in order to regulate technology effectively. It is evident that AI-based systems and services grab the attention of policymakers and researchers in light of recent regulatory efforts, like the EU AI Act and subsequent public interest technology initiatives focusing on AI (Züger & Asghari, 2023). While these initiatives have their merits, they end up narrowly focusing efforts on the latest trends in digitalisation. We argue that this approach leaves untouched the engineered environments in which digital services are produced, thereby undermining efforts to regulate AI and to ensure that AI-based services serve the public interest. Even when policymakers consider how digital services are produced, they assume that AI is captured by a few cloud companies (Cobbe, 2024; Vipra & West, 2023), when, in fact, AI is a product of these environments. If policymakers fail to recognise and tackle issues stemming from AI’s production environments, their focus on AI may be misguided. Accordingly, we argue that policy and research should shift their regulatory focus from AI to its production environments.

The White Paper sets out an initial high-level framework for states, policymakers, civil society, workers and others to dismantle Big Tech's concentrated power over digital ecosystems, and to encourage the emergence of a fair digital economy that is open, decentralised, democratic and serves the common good.

The Green Web Foundation tackles "How power consolidation of digital infrastructures threatens our democracies-and what we can do about it"

There is a growing trend for high-status original equipment manufacturers (OEMs) such as premium electronics manufacturers and premium carmakers to create and capture value through digital extensions of their products. However, these incumbents face disruptive threats from platforms offering substitutes for these digital extensions. The literature suggests that coopetition—the interplay of cooperation and competition—is a viable strategic response to this threat. However, we have a limited understanding of how high-status OEMs coopete with platforms to maintain their digital extensions' edge over time. We address this gap through a longitudinal case study of InnoCar, a premium European carmaker whose digital extensions—car-specific digital services that enhance the driving experience, such as real-time navigation and infotainment—were challenged by Google and Apple. In response, InnoCar pursued what we call the slipstream strategy, which consists of two phases with varying intensities of cooperation and competition. A high-status OEM first increases its cooperation with platforms at the expense of competition in order to establish shared demand-related complementary assets. Second, it focuses on competing with platforms on the quality of its digital extensions while keeping cooperation to a minimum. We develop a conceptual framework that specifies the slipstream strategy and provide boundary conditions for its application. Our paper contributes to research on coopetition with platforms.

Woodrow Hartzog - Industry will take everything it can in developing artificial intelligence (AI) systems. We will get used to it. This will be done for our benefit. Two of these things are true and one of them is a lie. It is critical that lawmakers identify them correctly. In this Essay, I argue that no matter how AI systems develop, if lawmakers do not address the dynamics of dangerous extraction, harmful normalization, and adversarial self-dealing, then AI systems will likely be used to do more harm than good.

Research on platform owners’ entry into complementary markets points in divergent directions. One strand of the literature reports a squeeze on post-entry complementor profits due to increased competition, while another observes positive effects as increased customer attention and innovation benefit the complementary market as a whole. In this research note, we seek to transcend these conflicting views by comparing the effects of the early and late timing of platform owners’ entry. We apply a difference-in-differences design to explore the drivers and effects of the timing of platform owners’ entry using data from three entries that Amazon made into its Alexa voice assistant’s complementary markets. Our findings suggest that early entry is driven by the motivation to boost the overall value creation of the complementary market, whereas late entry is driven by the motivation to capture value already created in a key complementary market. Importantly, our findings suggest that early entry, in contrast to late entry, creates substantial consumer attention that benefits complementors offering specialized functionality. In addition, the findings also suggest that complementors with more experience are more likely to benefit from the increased consumer attention. We contribute to platform research by showing that the timing of the platform owner’s entry matters in a way that can potentially reconcile conflicting findings regarding the consequences of platform owners’ entry into complementary markets.

There has been an explosion in uses of educational technology (EdTech) to support schools’ teaching, learning, assessment and administration. This article asks whether UK EdTech and data protection policies protect children's rights at school. It adopts a children's rights framework to explore how EdTech impacts children's rights to education, privacy and freedom from economic exploitation, taking Google Classroom as a case study. The research methods integrate legal research, interviews with UK data protection experts and education professionals working at various levels from national to local, and a socio-technical investigation of the flow of children's data through Google Classroom. The findings show that Google Classroom undermines children's privacy and data protection, potentially infringing children's other rights. However, they also show that regulation has impacted on Google's policy and practice. Specifically, we trace how various governments’ deployment of a range of legal arguments has enabled them to regulate Google's relationship with schools to improve its treatment of children's data. Although the UK government has not brought such actions, the data flow investigation shows that Google has also improved its protection of children's data in UK schools as a result of these international actions. Nonetheless, multiple problems remain, due both to Google's non-compliance with data protection regulations and schools’ practices of using Google Classroom. We conclude with a blueprint for the rights-respecting treatment of children's education data that identifies needed actions for the UK Department for Education, data protection authority, and industry, to mitigate against harmful practices and better support schools.

In recent years, Amazon, Microsoft, and Google have become three of the dominant developers of AI infrastructures and services. The increasing economic and political power of these companies over the data, computing infrastructures, and AI expertise that play a central role in the development of contemporary AI technologies has led to major concerns among academic researchers, critical commentators, and policymakers addressing their market and monopoly power. Picking up on such macro-level political-economic analyses, this paper more specifically investigates the micro-material ways infrastructural power in AI is operated through the respective cloud AI infrastructures and services developed by their cloud platforms: AWS, Microsoft Azure, and Google Cloud. Through an empirical analysis of their evolutionary trajectories in the context of AI between January 2017 and April 2021, this paper argues that these cloud platforms attempt to exercise infrastructural power in three significant ways: through vertical integration, their complementary innovation, and the power of abstraction. Each dynamic is strategically mobilised to strengthen these platforms’ dominant position at the forefront of AI development and implementation. This complicates the critical evaluation and regulation of AI technologies by public authorities. At the same time, these forms of infrastructural power in the cloud provide Amazon, Microsoft, and Google with leverage to set the conditions of possibility for future AI production and deployment.

Will policymakers permit an open-source future?

some proposals for the Global Digital Compact (GDC) can be read to mandate more centralized governance. If the final document contains such language, we believe it will be detrimental to not only the Internet and the Web, but also to the world’s economies and societies.

Panelists from the Atlantic Council give a bit of a primer on Internet Governance history, how things changed with the Snowden revelations, the recent push for multilateralism, and how China sees the Global Digital Compact as a way to gain more leverage. (Also some discussion of the Digital Silk Road initiative)

Panelists from the Atlantic Council give a bit of a primer on Internet Governance history, how things changed with the Snowden revelations, the recent push for multilateralism, and how China sees the Global Digital Compact as a way to gain more leverage. (Also some discussion of the Digital Silk Road initiative)

Is Google Search a monopoly with gatekeeping power? Reg-ulators from the US, UK, and Europe have argued that it isbased on the assumption that Google Search dominates themarket for horizontal (a.k.a. “general”) web search. Googledisputes this, claiming that competition extends to all verti-cal (a.k.a. “specialized”) search engines, and that under thismarket definition it does not have monopoly power.In this study we present the first analysis of Google Search’smarket share under vertical segmentation of online search.We leverage observational trace data collected from a panelof US residents that includes their web browsing history andcopies of the Google Search Engine Result Pages they wereshown. We observe that participants’ search sessions begin atGoogle greater than 50% of the time in 24 out of 30 verticalmarket segments (which comprise almost all of our partic-ipants’ searches). Our results inform the consequential andongoing debates about the market power of Google Searchand the conceptualization of online markets in general

Delineating a robust investment matrix in the AI start-up ecosystem from 1907 to 2024, this paper highlights key financiers like Sequoia Capital and Softbank. Dominant sectors receiving support include Biotechnology, Cloud Computing, and Generative AI, with a noticeable trend toward Web3 and NLP technologies. The network map emphasizes the dynamic interplay between established venture groups, accelerators, and government-backed entities, underscoring their vital role in nurturing innovation within this high-growth industry.