Review article of: A. Acker (2025). Archiving Machines. Technology and the Future of Memory Institutions, MIT Press, Cambridge (Ms)-London, 248 pp. To download as open source: https://direct.mit.edu/books/oa-monograph/6055/Archiving-MachinesFrom-Punch-Cards-to-Platforms.

Voor een Nederlandse vertaling klik hier.

Dr G.J. van Bussel

The author

Amelia Acker is an associate professor in the School of Communication and Information at Rutgers University. Her scholarship focuses on the generation, standardization, and preservation of digital information. Her research centres on three pillars: data technologies (the tools used to generate and capture information), digital archives (the management of digital data), and information in­frastruc­tures (the systems that facilitate long-term preservation and serve as repositories for cultural memory). Prior to her appointment at Rutgers, she served as Director of the Critical Data Studies Lab at the University of Texas at Austin. She led interdisciplinary research initiatives concerning the societal impact of ubiquitous mobile technology, public and academic understanding of data systems, and frameworks that enable information to remain searchable and functional over time. ¹

The key thesis

Archiving Machines. From Punch Cards to Platforms offers an historical overview of storage technologies and processes, tracing their evolution from the data banks of the 1960s to the emergence of contemporary digital platforms. At the heart of Acker’s argument is the assertion that storage systems function as ‘archiving machines’ that actively shape memory, regulate access, and distribute power. Grounded in Science and Technology Studies and Infrastructure Theory, but also using insights from Media Studies, she contends that these machines define the boundaries of cultural memory, determine its control, and define its organization.

As she clarifies, these machines are ‘not about traditional archives found in institutions but rather about how accessing knowledge has transformed as computational pro­cesses for storing and managing data developed, thus putting pressure on concepts of archives, archive, and archiving. In other words, this book explores how ‘archive’ became a verb with the rise of data-­driven recordkeeping to unpack the stakes of archiving machines’ (p. 21). Her work emphasizes the systemic impact of digital and platform-based infrastructures on collective memory, in particular the ways in which automation and corporate platforms have assumed roles once reserved for archival institutions and professional archivists.

She states ‘how firms who play an infrastructural role in data archives have become ­ today’s archons – ­ those that govern and assert power through the functional sovereignty of data archives’ (p. 11).Acker demonstrates how storage technologies have not only redefined what is preserved but have also reshaped conceptions of memory, ownership, and institutional accountability. Her thesis – that archiving machines shape memory and produce distinct regimes of access and data sovereignty – provides historical context for the challenges in digital stewardship that she identifies in the book. ²

Chapter overview

Acker elaborates on her thesis in Chapter 1, where she explores the conceptual transformation of ‘archive’ between the 1950s-1980s from a noun to a verb and the implications of this grammatical shift. The transformation from ‘archive’ (place) into ‘to archive’ and ‘archiving’ (process) was largely due to the emergence and development of computing. ³ This transformation signals a shift in the management of memory, evidence, and knowl­edge, now (largely) dominated by platforms and corporate infrastructures. The five subsequent chapters develop this argument chronologically, each examining a pivotal moment in the technological evolution of data storage.

Chapter 2 traces the transfer of information processing to machines, which led to the development of standardized systems for representing data. This established enduring patterns of control and initiated a critical ‘distancing’ between data and its creators, a trend that intensified in later technological iterations. Acker uses the example of the US National Data Centre to illustrate ‘a pivotal moment in twentieth-century data management’ (p. 24): the shift from manual data transcription to machine-readable data due to new storage technologies combined with machines that could process data on a large scale.

Chapter 3 analyses the emergence of programmable data storage, demonstrating how this transition enabled unprecedented forms of data manipulation. It marked the beginning of the concentration of technical expertise and control within specialized roles, creating dependencies that would later define platform-based relationships. It also initiated the abstraction of data from user control, a process that accelerated with the shift from file-based computing to application-centric and cloud-based models, as discussed in Chapter 4. Here, control migrated to networked services, where platform companies mediated access through proprietary applications rather than direct file manipulation.

Chapter 5 explores the role of mobile computing and telecommunications infrastructures in data creation and storage, describing them as ‘a peculiar type of archiving machine’ (p. 114). Acker highlights how these systems rely on ‘the massive amounts of automated metadata created as part of our everyday human activities …, creating both closeness and distance between data and people’ (p. 114). This process facilitates the ongoing abstraction of data from user control.

Chapter 6 examines the access regimes of platform companies and the mechanisms through which they govern data use. Acker argues that these regimes ‘fall short of traditional transparency and au­then­tic goals of archives to preserve trustworthy evidence’ (p. 162). She compares these access regimes with historical models of archival power and compares them with ‘archives as prisons—­ where information is controlled, and power is exerted through limiting access’ (p. 162).

In the Epilogue, Acker concludes that datafication, following the historical development of distancing data from users led to a situation in which ‘new power formations generate inequalities by controlling access to digital cultural memory’ (p. 172). Access to data is asymmetrical: platform companies control what is preserved, how it is used, and who can retrieve it, often prioritizing commercial uses over transparency or public good. This process perpetuates ‘the production of more data for further extraction accumulation for machines to access’ (p. 171). The rise of generative AI-models exposes how user data, once collected, are never truly ‘at rest’. It is continuously processed, repurposed, and fed into algorithms, creating a feedback loop: more data, more extraction, more platform power, and more obscurity of the origins and contexts of the data. Archiving machines, she asserts, are central to understanding what ‘this model of accumulation of data from­ human action means for the ways in which we access evidence, digital culture, and memory now as we see the datafication of all possi­ble­ human activities pro­cessed through platforms’ (p. 170).

Acker’s analysis is largely based on Philip Agre’s theory of ‘grammars of action’, which suggests that organizational systems act as structured environments that impose logic on human behaviour. These ‘grammars’ represent the patterns embedded within systems that enable and constrain how individuals interact with data and tools. ⁴ Acker illustrates this concept with the revealing example of the absence of a ‘save’ button in online word processors (p. 1). Although this design choice is presented as a convenience, facilitated by automated cloud synchronization, it subtly distances users from their data, both psychologically and technically. It diminishes individual agency over one’s work, shifting control from the user to the infrastructures of platform owners.

Strengths

Acker’s central thesis is original and intellectually stimulating. It invites readers to reconsider how storage systems influence memory and data sovereignty. Her interdisciplinary approach enriches her analysis, allowing her to explore the technical, social, and political dimensions of archiving machines.

The chronological structure of the book traces the evolution of archiving machines from data banks to cloud computing and mobile platforms. She provides a comprehensive historical narrative that contextualizes contemporary digital dilemmas. Each chapter focuses on a critical juncture in this evolution, demonstrating how technological shifts have reshaped data control, ownership, and access.

Acker excels in her analysis of power dynamics, particularly in how platform companies have appropriated functions once reserved for public institutions. Her discussion of platform sovereignty, whereby firms act as modern ‘archons’ governing data access, is particularly interesting. She illustrates how automation, proprietary infrastructures, and algorithmic gatekeeping generate novel control systems that frequently undermine transparency and user autonomy. Throughout, her prose remains accessible and engaging, balancing complex ideas with concrete examples.

Acker’s explorations of datafication, surveillance, and stewardship are relevant to our current, platform-dominated world. Her argument of how archiving machines produce inequalities – by controlling access to data – resonates with ongoing debates about data ethics, privacy, and the role of technology in society.

Her conclusions stimulate awareness. She calls for alternative models of data access that disrupt platform monopolies, suggesting that the ability to delete, disconnect, or store data offline is essential to counter the continuous extraction of digital data. She urges for public alternatives to commercial platforms, such as non-profit data cooperatives or open-source tools that empower users to manage their own data. She leaves readers with a critical challenge: How can we reclaim agency over digital memory? The future of archiving machines is not just a technical question but a political one, demanding collective action to reconfigure access, resist dispossession, and reimagine stewardship in the age of platform capitalism.

Reflections

Problematic Terminology

One of the analytical weaknesses of Archiving Machines is its inconsistent use of terminology. With the exception of archiving machines, Acker does not define her core concepts. This creates more than just occasional ambiguity. It undermines the precision of an argument that depends on distinctions between what is preserved, what is controlled, and what constitutes evidence. This issue goes beyond mere editorial inconsistency. It affects the book’s central conceptual claim.

‘Archiving Machines’

Acker defines ‘archiving machines’ as ‘apparatuses connected through layers of networks to support the saving and storage of information, ultimately preserving evidence, that then encode how it can be accessed’ (p. 10). While elegant, this definition is problematic in its breadth. Under this concept, Acker groups together technologies with radically different functions, purposes, and relationships to evidence: data banks, database management software, cloud storage platforms, and social media applications.

Data banks were designed as large-scale repositories for long-term preservation and access, functioning in close alignment with traditional archival principles. Databases, by contrast, are dynamic systems optimized for real-time processing and operational tasks, prioritizing efficiency over preservation. Social media platforms are algorithmically mediated commercial environments in which data is exploited for engagement and advertising rather than retained as evidence. Cloud infrastructure provides scalable, on-demand storage without preservative intent. Grouping all four under the same term renders the term archiving machines so expansive that it loses analytical force, particularly considering questions of evidence, memory, and institutional power.

Yet the problem is not only one of terminological imprecision. It is also a question of whether ‘archiving’ is the right conceptual frame at all. What Acker describes throughout her analysis is less a process of archiving than one of systematic ‘distancing’: the separation of users from their data through mechanisms such as automatic saving, proprietary formats, and algorithmic curation. These mechanisms erode users’ control, access, and understanding of their information, enabling platform companies to manage and exploit that data on their own terms. Acker uses the phrase ‘distancing techniques’ to describe this dynamic, and her own most revealing formulation – the ‘space between the data people generate and the cloud computing architectures that serve as utilities to store and manage data’ (p. 10) – suggests that this space, this distance, is the defining feature of what she is analysing. The examined systems may therefore be more precisely characterized as ‘distancing machines’, instruments of separation rather than of storage. This is a term Acker’s own analysis seems to reach for. It would have avoided lending an air of preservation, trust, and institutional neutrality to corporate platforms, qualities her own analysis demonstrates to be illusory.

Distinguishing between data banks (preservative and curated), databases (operational and dynamic), social media platforms (commercial and algorithmic), and cloud infrastructure (scalable and service-oriented) would have clarified the distinct consequences each carries for evidence, memory, and power, and would have enabled a more precise account of the different ways in which each distances users from their data.

Data, Records, and Documents

Acker uses the terms ‘data’, ‘records’, and ‘documents’ interchangeably (with ‘information’ also used many times). These terms are not synonymous. Records are created or received in the course of activity and retained as evidence of that activity. Documents are recorded information regardless of format or purpose. Data, in the strictest sense, refers to raw, uninterpreted material. These distinctions matter precisely because Acker’s argument turns on questions of what is being created, preserved, and controlled – and by whom. Treating them as a single, undifferentiated concept weakens her analysis of platform governance, user agency, and evidential integrity.

Data Archives

A related difficulty concerns the terms ‘data archive’ and ‘data archives’, which appear continuously in the book without ever being explicitly defined, an omission that presents a significant analytical challenge. Both terms are employed in professional practice with multiple interpretations, yet neither is explicitly defined in established archival or computing lexicons. The Multilingual Archival Terminology of the International Council on Archives, for instance, mentions data archives only indirectly, in its definition of a ‘machine-readable record’, which is described as being ‘organized in accordance with the principle of provenance as distinct from data archives’, implying that the latter are not organized in such a way. Acker’s thesis appears to align with this interpretation, yet she never makes this alignment explicit. ⁵

What she does instead is apply ‘data archive’ inconsistently across three distinct categories: institutional repositories adhering to traditional archival principles, such as the Roper Public Opinion Research Centre (p. 28) and university-based social science repositories (pp. 30–32), platform companies whose practices she elsewhere contrasts with those of archives, and large datasets (pp. 115-117) that cannot, by any reasonable definition, constitute a ‘place’ (as she interprets archives on p. 7). Given her assertion that a platform company ‘is not an archive’ (pp. 157–162), this introduces an inconsistency that undermines the precision of her argument.

Using clearer terminology would have made the argument much stronger, strengthening rather than weakening her argument about how platform companies misappropriate archival language.

Data Sovereignty, Preservation, and Memory

The same imprecision characterizes three further concepts that are central to Acker’s argument. ‘Data sovereignty’ shifts between custody of authentic records, ownership of intellectual property, and control over personal information without the distinctions being explained. ‘Preservation’ moves between bit-level integrity, long-term accessibility, and protection from corporate exploitation, concerns that are related but require different solutions. ‘Memory’ swings between individual memory, organizational memory, and cultural memory without adequate attention to the distinct archival challenges each presents. A teenager’s Instagram posts, a corporation’s internal communications, and a social movement’s Twitter archive all constitute ‘memory’ in some sense, but they raise entirely different questions about preservation responsibility, access rights, and archival value, questions that Acker’s undifferentiated use of the term leaves unaddressed.

The ‘Social’ or ‘Cultural Archive’

Finally, Acker’s analysis touches on a broader conceptual error that has become widespread in humanities scholarship: the characterization of aggregated social media content as a ‘social archive’ or ‘cultural archive. ⁶ While platform companies may offer research utility, they lack the integrity, transparency, and ethical rigour of archival preservation. This is not merely a scholarly concern. Facebook’s own admission in 2022 that it did not have ‘an adequate level of control and explainability’ over how its systems used data, confirms that these datasets are governed by commercial imperatives rather than preservative ones. ⁷ Platform companies resemble, in this respect, libraries more than archives. Libraries organize and provide access to materials according to institutional and commercial logic, but they do not claim, and cannot sustain, the evidential authority of archives. This tension is particularly evident in Acker’s use of the term ‘evidence’. She interprets it through a societal and epistemological lens, framing it as the raw material of history, knowledge, and identity, something preserved to document societal trends, behaviours, and events. However, platform companies do not treat their data as cultural evidence, and not even as ‘memory’. And given the way they handle data, it is questionable whether those datasets can be accessed in that way.

Acker’s stated aim is to expose how platform companies appropriate archival language to obscure their commercial motives. That is a powerful argument. But it is an argument that depends on maintaining clear distinctions between what archives are and what platform companies are. Those distinctions are precisely what her own inconsistent vocabulary puts at risk. The concept of ‘distancing machines’, latent in her own analysis, would have provided not only a more precise characterization of the systems she examines, but a terminological foundation consistent with the critical force of her argument.

Overgeneralization

Amelia Acker’s conceptualization of platforms as archiving machines is most persuasive when applied to individual users engaged in personal archiving, such as storing social media posts, personal documents, or audiovisual content. Had she centred her analysis on this context from the beginning, her argument would have been far more compelling.

However, the realm of archiving machines extends far beyond individual users, encompassing a far more complex landscape that includes not only traditional organizations (e.g. corporations, universities, government agencies) but also platform-dependent organizations, such as online-only businesses operating exclusively on platforms like Shopify, Amazon Marketplace, or WooCommerce.

Her analysis fails to distinguish between three different categories of platform users, each with unique levels of data sovereignty, control, and vulnerability:

• Individual users (e.g. a teenager on Instagram), who have no custodial control, are fully subject to platform governance, and are entirely dependent on the platform’s policies and algorithms. They have no legal or operational recourse to challenge data use or access restrictions.
• Traditional organizations (e.g. corporations using Salesforce or O365) have partial independence through contracts, local backups, hybrid systems, and legal leverage. They have the ability to enforce retention policies, pursue remedies for breaches, or migrate data to alternative systems. They operate with a degree of autonomy from platform dependency and have the option of mounting their own infrastructure or use an encryption technology independent of platform services.
• Online-only companies (e.g. Amazon Marketplace sellers) are legally organizations with responsibilities such as GDPR compliance, tax obligations, and customer data protection. However, they are operationally as dependent as individuals, with no independent infrastructure to fall back on. They also have no physical or digital alternatives, making them entirely reliant on the stability, policies, and algorithms of the platform.

By grouping these three categories under the same archiving machine framework, Acker overgeneralizes, obscuring the gradations of data sovereignty and control, the power dynamics at play, and the unique risks faced by online-only companies. Unlike traditional organizations, these entities are wholly reliant on the platform’s infrastructure, creating vulnerabilities that resemble those of individual users, but with far greater consequences such as legal liabilities, business continuity, and competitive survival.

The dependency of online-only companies on platforms becomes particularly evident in five critical scenarios, where they face greater risks than traditional organizations while sharing the same vulnerabilities as individual users: existential vendor lock-in, being subject to platform algorithms with no control or recourse, being fully exposed to platform policy changes with no alternative infrastructure to ensure compliance, secondary data use and monetization without consent, and being unable to fall back on local servers or alternative tools.

Acker’s framework fails to address how platforms reshape power dynamics for entities that are neither fully autonomous organizations nor helpless individuals, but rather a hybrid category with unique risks. Acker’s archiving machines framework overlooks this distinction, thereby ignoring complexity in the landscape of data sovereignty and control. A more nuanced approach would differentiate between the three categories of platform users, recognizing their unique vulnerabilities and power dynamics. Doing so would make her argument even more compelling than it already is.

Acker’s analysis misses an opportunity to fully illuminate how platforms reshape sovereignty, agency, and risk for a growing category of digital entities: those that are neither fully autonomous nor fully powerless but trapped in a precarious in-between. This oversight is particularly critical for understanding the future of digital commerce, labour, and governance, where platform dependency is not just an individual concern but a structural vulnerability for entire classes of organizations.

The ultimate ‘distancing machine’

Although Amelia Acker does not devote substantial attention to artificial intelligence (AI) – the term appears only briefly across four pages of Archiving Machines – her analysis nevertheless establishes the critical groundwork for understanding AI as an inevitable outcome of the systems she examines. The core themes that define AI’s development and deployment – data accumulation, algorithmic control, asymmetric access, and corporate sovereignty over data – are deeply embedded in her discussion. While she rarely uses the term ‘AI’ explicitly, her work anticipates its rise by uncovering the mechanisms that make it possible: the transformation of public and personal data into corporate-controlled collections and the systematic erosion of user agency over their own information.

In light of Acker’s stated aim to ‘chart the history of computing in the United States from 1960 to the present’ (p. 16), the exclusion of AI as an explicit archiving machine is noteworthy. If her project aims to trace the evolution of computing systems that distance users from their data, then AI, particularly in its contemporary generative forms, represents the most advanced iteration of this trend. The accumulation of data in corporate platforms, algorithmic mediation of access, and monopolization of information by platform companies are not merely precursors to AI but its foundational pillars. AI systems, particularly large language models and generative tools, depend entirely on the infrastructures that Acker criticizes – vast datasets scraped from platforms, exclusive control over data flows, and the systematic erosion of user agency.

Acker’s framework implies, without fully articulating, that AI constitutes the next phase of the archiving machine, one that does not merely store data but actively transforms it into new forms of corporate control. Large language models and other AI systems train on vast datasets scraped from the web, including platform-stored content that users may have believed they controlled but which, in reality, resided in corporate custody. This content then undergoes a further layer of extraction: processed, transformed into model weights, dissolved into probabilistic patterns, and reassembled into outputs that bear traces of the original but are entirely severed from the user’s context and custody. What emerges is content without context, knowledge without identifiable origin – the logical endpoint of the distancing dynamic Acker traces across six decades of computing history.

The omission of AI as an explicit focus is notable given this scope. Whether it reflects a strategic choice to concentrate on foundational systems, a temporal constraint, or an intentional delimitation of scope is not made explicit in the book. Whatever the reason, the absence is consequential. Given that her analysis so thoroughly exposes the infrastructures that make AI possible, the failure to examine it as the culminating expression of these trends leaves a meaningful gap in an otherwise comprehensive argument. The emergence of generative AI, in particular, suggests that the very terminology of the book deserves reconsideration: these systems are not primarily archiving machines but distancing machines in their most advanced form, dissolving the relationship between users, their data, and any recoverable trace of context or ownership.

Pressure of Concepts

Amelia Acker’s assertion that her work is ‘not about traditional archives found in institutions but rather about how accessing knowledge has transformed as computational processes for storing and managing data developed, thus putting pressure on concepts of archives, archive, and archiving’ (p. 21) serves as a deliberate framing of her argument. Her focus lies on platform companies, such as Amazon, Facebook, and Google, that mimic, distort, or outright replace the functions of traditional archives. These corporations accumulate, control, and exploit data, adopting the language of archival practice while failing to uphold the ethical and operational standards, practices, and principles that define archives: provenance, transparency, public access, and evidential integrity. By invoking these standards and practices, Acker implicitly contrasts corporate data practices with the ethical frameworks of traditional archives, exposing how platforms fall short of these standards.

Yet Acker’s argument, while compelling, overlooks a critical distinction: providing access to data or records is not the same as providing access to knowledge. Platforms do not facilitate knowledge; they deliver data that is distorted, fragmented, and algorithmically mediated. Knowledge is not a static dataset, but an abstraction constructed in the human mind through the interpretation of information. However, platforms actively manipulate the conditions under which this interpretation occurs. Through algorithmic curation, obscure processing, and the systematic stripping of context, they prevent users from forming independent understanding. Users are left with what amounts to a curated illusion: an interpretation of reality that aligns with the platform company’s interests, not their own.

Platform companies do not merely distort archival standards, practices, and principles. They exploit the gap between data and knowledge to control how information is interpreted. By shaping the conditions of access, platforms ensure that users cannot transform data into knowledge on their own terms. Instead, users receive a pre-packaged narrative, one that serves corporate interests rather than fostering genuine understanding. This is not just a failure of transparency; it is a systematic strategy of control, where platforms dictate not only what users see but also how they are allowed to understand it.

While Acker rightly critiques corporate data practices by contrasting them with archival standards, practices, and principles, such as provenance, transparency, evidence, and public access, she overlooks a fundamental truth: records are never neutral, objective, or complete, even within traditional archives. Archives do not preserve pristine or impartial records; instead, they manage records already shaped by human behaviour, organizational biases, and institutional priorities. They are inherently incomplete and distorted, whether through unauthorized deletions, retrospective alterations, deliberate design choices to exclude certain information, or selective appraisal processes that privilege some narratives over others. ⁸ However, there remains a critical distinction: records managed by archives are subject to professional standards, practices, and principles aimed at preserving context, ensuring accountability, and enabling public scrutiny. Platform companies are designed to serve operational, commercial, or strategic goals at the expense of archival standards, practices, and principles. This distinction reveals how platform companies exploit the rhetoric of archives to prioritize profit over ethical stewardship. Their datasets are proprietary assets, managed to maximize control and minimize accountability.

Acker’s assertion that computational processes exert ‘pressure on concepts of archives, archive, and archiving’ is only partially accurate. In truth, both traditional archives and platform companies manage distorted records and data, whether through human intervention, institutional biases, or technical processes. However, what distinguishes platform companies is not merely the presence of distortion but how they automate and scale these distortions on an unprecedented level, rendering them more systematic, pervasive, and far less accountable than institutional archives.

This shift represents a leap in the manipulation of information: where traditional archives may distort records through selective appraisal, retrospective alterations in metadata, or institutional priorities, platforms industrialize these distortions through algorithmic curation, obscure processing, and proprietary control. The result is a system where distortion is not an exception but a designed feature, one that erodes transparency, undermines provenance, and consolidates power in ways that traditional archives have never done.

Archival Theories

What makes reading this book perplexing is the absence of archival science theory from a work so deeply concerned with archives, data archives, storage, accountability, provenance, public access, and archiving machines. For all its engagement with core archival concepts, the bibliography tells a revealing story: just eighteen references to archival theoretical literature — and this count is generous in classifying some of them as archival theory – with only seven from 2015 onwards. Nowhere to be found are seminal contributions such as Frank Upward’s Records Continuum Theory, Luciana Duranti’s Digital Diplomatics and the InterPARES projects, or emerging frameworks such as Archive-as-Is, which – whatever its current position in the mainstream literature – addresses questions of situational and environmental context that are directly relevant to Acker’s concerns. Given that the book’s central concern is how platforms distort, exploit, and control data, this omission is especially puzzling: these are precisely the theoretical frameworks that could have sharpened her argument. ⁹

Archival theory, despite its obvious relevance to a book about archives, data archives, and archiving, is frequently perceived within adjacent fields as either too practice-bound or too narrowly fixated on institutional recordkeeping to offer much analytical depth on infrastructural ‘archives’. Acker may well have concluded – understandably from within her own disciplinary tradition – that the frameworks she chose were better suited to her purposes. Her intellectual home is Information Studies, Science and Technology Studies, and Media Studies. This shapes how she frames her arguments. Disciplinary orientation is not a flaw; it is a choice. But it carries a cost. Archival science has developed theoretical frameworks that speak to the questions she is asking – data sovereignty, accountability, provenance, the ethics of recordkeeping – and by leaving them aside, her analysis, though often insightful, sacrifices the theoretical precision and depth that engaging with that tradition would have afforded.

When a scholar sets out to examine the politics of how data, including records, are created, stored, migrated, and controlled, and does so without once engaging modern archival theory, something structural is at work. It suggests that the boundaries separating Information Studies, Media Studies, and Archival Science remain surprisingly high, even when all three fields are circling the same objects. The walls, it seems, have not come down nearly as far as the vocabulary might imply.

Records Continuum Theory, to begin with, would have allowed Acker to map precisely how platforms disrupt the continuum of records (Create, Capture, Organize, and Pluralize). Framing them as such would have shown how they transform records from evidence of action into raw material for corporate extraction, offering a systematic account of how accountability erodes and users are alienated from their own data, rather than an accumulation of telling but disconnected observations.

Digital Diplomatics would have served a different but equally valuable purpose: enabling Acker to diagnose platforms as fundamentally untrustworthy systems. Applying its criteria of authenticity, reliability, integrity, and usability, she could have demonstrated how platform companies violate these standards at every turn, through algorithmic curation, opaque processing, and the deliberate restriction of access to raw data. Grounding the argument in these established principles would have given it an authority harder to dismiss.

The Archive-as-Is framework would have added a third dimension. With its insistence on preserving the original situational and environmental context of records (and all other context added to the records during their existence), it would have provided a lens for examining how platforms strip data of its provenance and meaning, repurposing what belongs to users for corporate gain. It would also have enabled her to evaluate the value of the data archives offered by platform companies as either a ‘social’ or a ‘cultural’ archive. ¹⁰

Taken together, engaging with archival theory would have made her argument both more precise and more actionable, moving Archiving Machines toward something closer to a manifesto for data sovereignty.

Conclusion

Archiving Machines. From Punch Cards to Platforms is an ambitious, intellectually stimulating, and often illuminating work that challenges conventional notions of how storage technologies shape memory, access, and power. Its central argument is both original and urgently timely. Her chronological account of how digital infrastructures have progressively distanced users from their data is compellingly constructed, and her critique of platform sovereignty resonates powerfully in an era where the consequences of corporate data governance are increasingly impossible to ignore. The book’s greatest strengths lie in its interdisciplinary reach, its accessible prose, and its unflinching exposure of the political economy of digital infrastructure.

Yet, for all its merits, Archiving Machines falls short of its full potential in ways that matter. The terminological imprecision that pervades its core concepts is more than an editorial issue; it is an analytical one. These ambiguities blur critical distinctions that are essential to the very argument Acker seeks to advance. More significantly, her failure to engage with archival science – the field whose intellectual tradition most directly addresses her subject – leaves her reinventing arguments that already exist, rather than leveraging their precision.

Equally notable is the book’s conspicuous silence on artificial intelligence, which, as the culminating expression of the trends Acker describes, would have been a natural and necessary extension of her historical narrative. The absence of AI as a focal point feels like a missed opportunity, especially given the book’s own logic, which cries out for this final piece of the puzzle.

These limitations, however, do not diminish the book’s considerable merits. Readers from Information Studies, Media Studies, Science and Technology Studies, and the Digital Humanities will find much here to engage with, debate, and build upon. For Records Management, Digital Preservation, and Archival Science experts, the book may prove both frustrating and energizing: frustrating in its disciplinary blind spots but energizing in its demonstration that the questions archivists have long grappled with are now pressing concerns for a broader public. The verdict is clear: read it critically, with an awareness of its conceptual limitations. Archiving Machines delivers a serious, searching, and often provocative contribution to a debate that is only growing in urgency. And that alone makes it more than worthy of attention.

---

1. Online sources, retrieved March 1, from: https://sci.rutgers.edu/acker-amelia and https://www.ameliaacker.com/. ↩︎
2. The concept of ‘functional sovereignty’ was originally introduced by Riphagen in the context of international law, where it denoted the exercise of sovereign-like authority beyond traditional territorial boundaries. Acker draws on Frank Pasquale’s subsequent application of the concept to digital governance, in which he argues that large platform companies have come to exercise a form of functional sovereignty over significant domains of online life, setting rules, adjudicating disputes, and regulating behaviour in ways that parallel, and increasingly rival, the functions of the state. The result, as Pasquale contends, is a new digital political economy in which corporate power progressively displaces democratic control. Where territorial sovereignty is grounded in geography and law, functional sovereignty is grounded in infrastructure and market dominance, a distinction that lies at the heart of Acker’s critique of platform governance. W. Riphagen (1975). ‘Some reflections on ‘functional sovereignty’, Netherlands Yearbook of International Law, vol. 6, pp. 121-165 and F. Pasquale (2017). ‘From territorial to functional sovereignty. The case of Amazon’, Law and Political Economy (LPE) Project, Vol. 6. Online source, retrieved January 15, 2026, from: https://lpeproject.org/blog/from-territorial-to-functional-sovereignty-the-case-of-amazon/. ↩︎
3. Acker’s dating of the shift from the noun ‘archive’ to the verb ‘to archive’ is not entirely accurate. In 2019, digital preservation specialist Trevor Owens cited evidence from Merriam-Webster that the verb form appeared as early as 1831, supplementing this with earlier examples from 1816 and 1823. These findings suggest that the history of ‘to archive’ as a verb is considerably older and more complex than Acker implies. While its widespread adoption did accelerate in the 1960s (when it became closely associated with data processing and computing ) its origins are not solely tied to the digital era. The verb emerged in administrative, bureaucratic, and scholarly contexts long before its later association with digital systems. Its equivalents in other European languages (‘archiver’ in French, ‘archiveren’ in Dutch, ‘archivieren’ in German) appeared in the nineteenth century and disseminated from there into other archival traditions, suggesting that the conceptual roots of the verb lie in nineteenth-century administrative and bureaucratic practice rather than in computing. This picture is further nuanced by Luciana Duranti, who notes that the term ‘archiving’, as used by computer scientists to refer to saving records to a system, entered the archival world in the 1990s carrying two distinct meanings: the transfer of organizational records to an archival repository for long-term preservation, and the creation of accumulations of documentary materials related to people’s lives or social events. The verb’s trajectory is therefore not a single shift but a layered one, moving across professional cultures and acquiring new meanings at each point of transfer, a complexity that Acker’s account, in its focus on computing, tends to flatten. T. Owens (2019). ‘When and how did ‘archive’ become a verb?’. Online source, retrieved January 17, 2026, from: https://trevorowens.org/2019/04/28/when-and-how-did-archive-become-a-verb/ and L. Duranti (2018). ‘Archiving’, C. Lury, R. Fensham, A. Heller-Nicholas, S. Lammes, A. Last, M. Michael, and E. Uprichard (eds.), Routledge Handbook of Interdisciplinary Research Methods, Routledge, Section 2, Chapter 3, pp. 95-98. ↩︎
4. P.E. Agre (1994). ‘Surveillance and capture. Two models of privacy’, The information society, Vol. 10, No. 2, pp. 101-127. ↩︎
5. See: International Council on Archives. Multilingual Archival Terminology. Machine-readable record. Online source, retrieved February 9, 2026, from: https://www.ciscra.org/mat/mat/term/230/441; Society of American Archivists (2006-2026). Dictionary of Archives Terminology. Online source, retrieved February 9, 2026, from:https://dictionary.archivists.org/index.html; Free Online Dictionary of Computing (Foldoc) (1985-2025). Online source, retrieved February 9, 2026, from: https://foldoc.org/; British Computing Society (2016). BCS Glossary of Computing, Swindon (14^th edition). ↩︎
6. M. Manoff (2004). ‘Theories of the archive from across the disciplines’, Portal. Libraries and the Academy, Vol. 4,nr. 1, pp. 9-25, p. 11. ↩︎
7. ‘We do not have an adequate level of control and explainability over how our systems use data, and thus we can’t confidently make controlled policy changes or external commitments such as ‘we will not use X data for Y purpose.’ And yet, this is exactly what regulators expect us to do, increasing our risk of mistakes and misrepresentation.’ See: L. Franceschi-Bicchierai (2022). ‘Facebook doesn’t know what it does with your data, or where it goes. Leaked document’, Vice, April 26. Online source, retrieved February 7, from: https://www.vice.com/en/article/facebook-doesnt-know-what-it-does-with-your-data-or-where-it-goes/. For the leaked document: https://www.documentcloud.org/documents/21716382-facebook-data-lineage-internal-document/. Acker quotes this document in note 65 of chapter 6. ↩︎
8. G.J. van Bussel (2020). A Sound of Silence. Organizational Behaviour and Enterprise Information Management. Papers on Information and Archival Studies, I, Van Bussel Document Services, Helmond, pp. 89-92. ↩︎
9. F. Upward (1996). ‘Structuring the records continuum, part one. Postcustodial principles and properties’, Archives and Manuscripts, Vol. 24, No. 2, pp. 268-285; F. Upward (1997). ‘Structuring the records continuum, part two. Structuration theory and recordkeeping’, Archives and Manuscripts, Vol. 25, No. 1, pp. 10-35; L. Duranti (1998). Diplomatics. New uses for an old science, Lanham and London, The Scarecrow Press. For the InterPARES projects see: https://www.interpares.org/; and G.J. van Bussel (2017). ‘The theoretical framework of the ‘Archive-as-Is’. An organization-oriented view on archives. Part I. Setting the Stage: enterprise information management and archival theories. Part II. An exploration of the ‘archive-as-Is’ framework’, F. Smit, A. Glaudemans, and R. Jonker (eds.), Archives in Liquid Times Stichting Archiefpublicaties, ‘s-Gravenhage, pp. 16-41, 42-71. ↩︎
10. G.J. van Bussel (2022). ‘Determining the value of a Digital Archive. The Framework for the ‘Archive–As–Is”, A. Öhrberg, T. Berndtsson, O. Fischer, and A. Mattsson (eds.), From Dust to Dawn. Archival Studies after the Archival Turn, University of Uppsala, Studia Rhetorica Upsaliensia/Uppsala Rhetorical Studies, Vol. 8, pp. 56-101. ↩︎