​
DHQ: Digital Humanities Quarterly
Editorial

From Digitization to AI Integration: Accessing Digital Archives with Insights from I Trust AI

Jessica Bushey <jessica_dot_bushey_at_sjsu_dot_edu>, , School of Information, College of Information, Data & Society San José State University, California, USA ORCID logo 

DOI: pending

Abstract

This article considers how artificial intelligence (AI) tools are reshaping access to digital archives. While AI offers new possibilities for metadata extraction, transcription, image analysis, and large-scale discovery, its use in archival contexts raises critical questions about trustworthiness, ethics, and professional capacity. Drawing on archival and digital humanities scholarship and recent research initiatives, particularly the I Trust AI Project, the article identifies three central challenges: protecting the trustworthiness of records, ensuring that AI adoption aligns with ethical frameworks that resist extractive practices, and overcoming the AI literacy gap within the archival profession. It argues that archivists must play a central role in shaping AI integration so that access systems remain accountable to rights, relationships, and community needs.

I. Introduction

If you can’t find it online, then it doesn’t exist.
True or False? For researchers and the public, digital presence increasingly signals cultural existence. For archivists, who know that roughly seventy-five percent of historical records remain offline, the claim is demonstrably false.
This tension between perception and archival reality frames the central question of this article: how can AI integration reshape archival access in ways that enhance discoverability, while preserving trustworthiness and ethical stewardship? As a researcher with the I Trust AI Project and an archival educator at San José State University, I bring both research and pedagogical perspectives to this question. This article discusses original research findings and situates them within archival and digital humanities debates.
In the digital age, accessibility of archives is determined by the discoverability of digitized and born-digital records online and their availability for analysis and reuse [Friedwald et al. 2024]; [Jaillant 2022a]. For digital humanists, access to digital archives determines not only which records can be discovered, but which records can be analyzed at scale using computational methods. When archival materials remain offline, or when digitized and born-digital images are not prepared for computational use, they may be effectively absent from the datasets available to digital humanities researchers [Colavizza and Jaillant 2026]. The result is an incomplete representation of archival holdings that can influence the types of research questions asked, the methods used and the interpretation of findings.
Archival institutions and organizations rely on technologies that support and provide online access to digital archives; however, the decision to select records for online access is determined by archivists and interested parties based on institutional policies, legal requirements, ethical frameworks, donor agreements, and cultural protocols [Hirtle, Hudson, and Kenyon 2009]; [Dryden 2014]; [Dryden 2017]; [United Nations General Assembly 2007]. Procedures for digitizing archives and processing born-digital records for access are not only labor-intensive but require extensive consultation with contributors, interested parties, and affected groups to determine the appropriate level of access and degree of reuse. The introduction of AI-powered multimodal platforms, such as ChatGPT, is changing how people expect to interact with information everywhere, including archives and cultural heritage organizations. The potential of AI tools and algorithms to support archival functions, particularly activities aimed at increasing access and reuse of digital archives, is being explored by the archives and records management field [Colavizza et al. 2021]; [Mannheimer et al. 2024] and the humanities [Jaillant 2024]; [Jaillant 2023a]; [Jaillant 2022a]; [Jaillant 2022b]; [Jaillant 2022c].
The impetus for this article grew out of responses to a presentation on harnessing machine learning (ML) for archival description and community sharing of digital archives at the Crossing Fonds Symposium (2024). The subsequent discussion underscored strong interest in the potential of AI to advance discovery of archival holdings and enhance access platforms, while also raising concerns regarding its risks, particularly in relation to ethics, trust, and requisite technical competencies. Those conversations identify a need for more than technical demonstrations; archivists want to critically evaluate AI tools according to organizational needs, legal requirements, and ethical frameworks. This article responds to that call by situating recent archival and humanities research, specifically I Trust AI studies, that are relevant to access.
AI offers promising solutions for both back-end productivity (i.e., metadata extraction, object recognition, transcription) and front-end discovery and analysis (i.e., distant viewing, image clustering, cross-collection analysis). But enthusiasm for AI’s ability to curate large volumes of data is tempered by the complexity and opacity of AI systems, which pose two major challenges for archives: first, safeguarding the trustworthiness (i.e., accuracy, reliability and authenticity) of records, ensuring that the evidentiary, legal, and historical value of archives is protected throughout processes of digitization and datafication; and second, ensuring that technological adoption aligns with critical ethical frameworks that reject colonial and extractive approaches in favour of social justice, consent, and collective stewardship. A third challenge is the steep AI knowledge and skills barrier for records managers and archivists.
While AI engineers hold technical expertise, archivists possess the theoretical and ethical expertise necessary to guide responsible AI implementation in recordkeeping environments. Bridging this divide requires AI literacy within the records management and archives profession and interdisciplinary partnerships with both humanities researchers and AI developers. Only through such collaboration can the archival community integrate AI into archival functions and access platforms in ways that are ethically sound, theoretically grounded and responsive to the needs of diverse communities.

II. Background Literature and Concepts

This section presents key concepts that are relevant to understanding how AI reshapes access to digital archives. It draws on recordkeeping standards, archival and humanities scholarship, and governance practices as a foundation for the discussion of recent I Trust AI findings.

Digitization and Datafication

Digitization refers to the process of creating a digital representation of a physical object and its associated information. Methods include imaging (scanning or direct capture) and converting analog signals (audiovisual media). The purposes are twofold: preserving vulnerable media and producing digital surrogates for access [FADGI 2023]. Over the past quarter century, government agencies have pursued paperless operations, while cultural heritage organizations have uploaded digitized surrogates into collections management systems (CMS). Guided by standards such as FADGI, ISO, and NISO, archivists digitize records in all formats and add descriptive, administrative, and rights metadata to protect their trustworthiness (i.e., accuracy, reliability, authenticity).
Early digitization efforts were shaped by copyright and privacy laws [Dryden 2017], but digitization quickly became a democratizing force. Digitization programs and projects were framed by the assumption that access was a right, designed to attract online audiences and increase discoverability of hidden holdings. This shift disrupted traditional methods of controlling reproduction and use, and the rise of search engines further encouraged multimedia, social, and frictionless models of access — a precedent for how AI is now reshaping user expectations.
Digitizing and presenting archives online also created new responsibilities for archives and records professionals. The balance between rights of creators and subjects with increasing demands for open access and reuse, especially regarding materials by and about Indigenous and historically marginalized communities. As the digital environment blurred provenance and ownership, issues of data sovereignty became more urgent. [Moss, Thomas, and Gollins 2018] argue that digitization transforms archives from systems defined by provenance and original order into reconfigurable datasets. The authors assert that “[d]atafication is ubiquitous” [Moss, Thomas, and Gollins 2018, 131], requiring archivists to engage with computational methods alongside historians who now rely on distant reading and algorithmic analysis. Their work underscores the stakes of AI integration, that without careful design, computational tools risk severing archival documents from their context of creation and use.

Computational Archival Science (CAS)

The demand for computational access to digital archives led to the emergence of Computational Archival Science (CAS) in 2016, introduced by a multidisciplinary team of scholars [Marciano et al. 2018]. The authors recognize that existing archival methodologies and tools are insufficient in the age of big data and that computational methods derived from data and computer science are needed; however, there is a divide between computer science’s focus on data and archival science’s emphasis on aggregates of records, provenance, and context.
The International Research Collaboration Network in Computational Archival Science (IRCN-CAS) was launched in 2019, further developing research capacity. The momentum has continued with the development of CAS curriculum for graduate programs in library, archives, and information studies [Robinson 2014]. Scholars have also emphasized CAS as a framework for interdisciplinary collaboration with digital humanities [Hedges et al. 2022]. [Mordell 2019] critically examines the implications of treating archives as data, noting that the process reshapes and reframes both content and context. The author stresses that CAS must actively include Indigenous and marginalized communities as collaborators and project leads, not just subjects [Mordell 2019, 158]. These critiques position CAS as a necessary, but ethically contingent, foundation for integrating AI into archival access.
The recent proliferation of AI tools and AI-generated content (e.g., textual records, images, videos) introduces new layers of complexity for archivists when approaching copyright, privacy, data sovereignty, and cultural protocols. Navigating these complexities requires transparency in how systems are designed, trained, and deployed, and collaboration with partners, collaborators, interested parties, and source communities. An essential part of that transparency is acknowledging the political, historical, and cultural forces that shape both archives and the technologies used to acquire, manage, and access them. These forces form the foundation for the archival ethics that must guide AI integration as discussed in the next section.

Archival Ethics

The digital turn expanded the reach of archives and invited new audiences to discover and use archival holdings. Digitization projects, widely funded across G+LAMs, were often framed as democratizing access by removing physical and geographic barriers. As infrastructure for hosting digital collections online became more available and affordable, community archives and historically underserved groups increasingly took active roles in digitization and sharing archives online [Bastian and Flinn 2020]; [Caswell and Cifor 2016]; [Benoit and Eveleigh 2019].
[Benoit and Eveleigh 2019] highlight how early participatory projects used Web 2.0 platforms, such as Flickr commons, to engage virtual users, harness crowdsourcing knowledge, and facilitate collaborative tagging. By opening archival processes to new contributors, participatory archives advanced a shift toward shared stewardship and away from institutional gatekeeping. The authors reflect on the increase in community-based collecting and sharing as part of the larger growth of social media and online platforms: “The participatory archive movement can be seen as an extension of postmodernism through an attempt to better integrate the user perspective within archival processes” [Benoit and Eveleigh 2019, Ch.1, p.6]. Earlier, [Cook 2013] similarly called for archivists to become facilitators and mentors, listening to Indigenous Peoples and valuing non-Eurocentric storytelling traditions [Cook 2013, 114]. These shifts align with broader archival debates on power, representation, and social justice [Gilliland 2011]; [Duff et al. 2013]; [Caswell 2024]; [Punzalan and Caswell 2016].
These shifts in archival theory and practice are particularly relevant to Indigenous records and data, where expanding digital access and computational reuse raise questions of authority, sovereignty, and community governance. The CARE (Collective Benefit, Authority to Control, Responsibility, Ethics) Principles for Indigenous Data Governance respond to open-data practices that have overlooked Indigenous sovereignty [Carroll et al. 2020]. CARE provides a roadmap for ethical data management grounded in Indigenous knowledge systems, an approach made more urgent by reports of cultural heritage collections being scraped for AI training [Weinberg 2025]. By weaving together governance, ethics, and technology, CARE demonstrates how access frameworks must respect rights and responsibilities alongside innovation.
Informed by their research on human rights, social justice and community archives, [Caswell and Cifor 2016]; [Caswell and Cifor 2019]; [Caswell and Cifor 2021] extend archival ethics through a feminist ethics of care. They frame archivists as caregivers “…tied to the records’ creators, subjects, users and communities through relationships of mutual affective responsibility based on radical empathy” [Caswell and Cifor 2019, 160]. In this model, digitization is not a one-time technical act but part of ongoing ethical engagement that seeks to repair structural inequities rather than reproduce them. Their insistence that “empathy is radical” when centered on those most oppressed [Caswell and Cifor 2021, 4] reinforces the need for archivists to foreground equity in digital contexts.
Taken together, these frameworks—participatory, CARE and feminist care ethics—underscore that access is never ethically neutral. Policies, legal restrictions, cultural protocols, and technological infrastructures shape how records are interpreted and used. For AI adoption, they serve as essential guardrails, ensuring that system design, data governance, and collaboration remain accountable to communities and that AI contributes to, rather than undermines, the trustworthiness of archives.

Online Digital Archives

Shifting archival theory and practice to embrace participatory approaches has found expression in technological advances. The development of content management systems (CMS) for archives and cultural heritage organizations, such as the International Council on Archives’ (ICA) Access to Memory (AtoM), provided standards-based solutions for streamlining archival description workflows and launching collections online. The ICA-AtoM Project (launched in 2008) aimed to raise awareness of archives, promote ICA descriptive standards and build professional capacity [Van Garderen 2009]. While reinforcing certain political and cultural structures, AtoM’s open-source flexibility allowed adaption to accommodate different levels of privacy and relationships between entities. A notable example is the National Centre for Truth and Reconciliation in Canada, which developed a customized version of AtoM as the digital architecture for its archives, designed as “a powerful agent… allowing visitors to understand the history of Canada and the mistreatment of Indigenous Peoples” [National Centre for Truth and Reconciliation (n.d.)].
The Mukurtu Wumpurranrni-kari Archive, launched in 2007, reimagined CMS through Indigenous knowledge models that foreground social relationships over collections [Christen 2019]. As a collaborative partnership between the Warumungu community of Australia and the Center for Digital Scholarship and Curation at Washington State University, Mukurtu CMS incorporates participatory features, such as protocol-based access, community commentary, linguistic dictionaries, and Traditional Knowledge (TK) Labels. In contrast to Western models of open access or copyright-driven restrictions, Mukurtu CMS protocols enable Indigenous communities to govern circulation, reuse, and repatriation of cultural materials [Rodriguez 2018]; [Christen 2019]. Sustained growth of Mukurtu CMS is supported by a network of Hubs offering regional support to communities and organizations developing online digital archives. Now used in over 800 communities, Mukurtu CMS demonstrates how community-driven governance can reshape technical infrastructures, offering lessons for AI adoption that must also respect local values and authority.
Collaborations among communities, technologists, archivists, and digital humanists have also driven innovation in user interfaces. The longstanding tension between traditional finding aids and Google-style search has led to experimentation with “generous interfaces” [Whitelaw 2012]. These visually engaging platforms allow users to browse and interpret digital collections in exploratory ways, anticipating the next phase of computational tools where digital humanities researchers employ algorithms to generate new interpretations across collections [Whitelaw 2012]. Such experiments highlight both the possibilities and potential challenges of integrating AI in archival access platforms.
The COVID-19 pandemic (2020-2023) accelerated the digitization of G+LAM holdings and intensified the collection of born-digital communications shared on social media platforms. Physical isolation, coupled with increased reliance on digital platforms for information, entertainment, and emotional connection, prompted organizations to explore social media as both a site of outreach and a source of acquisition [Bushey 2023]. Rapid-response collecting drew attention to community-centered initiatives such as Documenting the Now, which advocated for ethical collection, use, and preservation of social media content related to historically significant events [Jules, Summers, and Mitchell 2018]. A central lesson from this project is recognizing the risk of accessing social media content for research and archival purposes, “without having to participate in equitable engagement as a way to understand [creators’] lives, communities, or concerns” [Jules, Summers, and Mitchell 2018, 3]. DocNow’s open-source tools introduced mechanisms for obtaining creator consent, embodying participatory principles. However, Twitter’s 2022 policy changes under new ownership broke these tools, illustrating the vulnerability of relying on commercial platforms [Summers 2023]. [Acker and Kriesberg 2020] identify the preservation and access challenges created when social media platforms and their application programming interfaces (APIs) determine how data can be extracted and reused. Approaches to accessing and downloading image collections and associated social media content by users, archivists, and researchers are dependent on changing terms of service and access policies. Social media are not default archives; their content is volatile, corporately controlled, and increasingly harvested as training data for large language models (LLM).
As digital archives grow and diversify, archivists and researchers must remain committed to preserving the contextual relationships between records and the people and events that create them. Designing access platforms and integrating computational tools, including AI, requires vigilance against extractive practices and a commitment to equitable engagement, context, and consent. These considerations establish the foundation for the following discussion of AI and archives research, where trustworthiness, ethics, and collaboration are paramount.

III. Discussion of AI and Archives Research Initiatives

The prior section outlined the conditions in which AI now enters archival practice. Archivists face the challenge of managing increasing volumes of digitized and born-digital records, including personal information, while researchers expect enhanced access across repositories, CMS platforms, and social media collections. At the same time, political, social and cultural change has driven archival theory toward recognizing critical ethical principles and Indigenous data sovereignty. Participatory frameworks and the rise of Computational Archival Science (CAS) both reflect and reinforce the increasing complexity of making trustworthy records accessible in the age of AI.
Against this backdrop, AI promises new capacities. Computer vision (CV) and machine learning (ML) offer tools for automating metadata creation, classifying images and enabling large-scale discovery across collections. Yet these promises bring new risks that must be assessed through the application of archival theory and critical ethics. The following sections survey recent research initiatives exploring archives and AI, before turning to early findings from the I Trust AI Project (2021-2026), a multinational, interdisciplinary effort to explore, design, and evaluate AI applications in records management and archival practice. Central to this initiative is the goal of building knowledge and tools through collaboration among academia, archival institutions, and industry while strengthening archival education and professional capacity [InterPARES Trust AI. (n.d.)].

Humanities-Led Archives and AI Initiatives

A significant body of research on AI and archives has been led by digital humanities scholars. Over the last five years, Lise Jaillant, Professor of Digital Cultural Heritage at Loughborough University, has directed multiple initiatives, including AURA (2020–2022), AEOLIAN (2021–2023), EyCon, LUSTRE (2022–2025), and LUSTRE/GLOW (2025-). These projects focus on enhancing access to cultural heritage archives and government records in digital form. Their activities range from convening workshops and webinars with G+LAM professionals to publishing case studies and open-access special journal issues. [1]
For archivists, these initiatives are valuable not only as technical demonstrations but as opportunities to understand how users and scholars engage with digital archives computationally. The resulting case studies highlight practical approaches to curating, interpreting and adding meaning to archives through AI. More importantly, they emphasize that user experience must be anticipated from the outset of AI integration in order to support scholarly needs, including distant reading and large-scale computational analysis [Jaillant and Rees 2023b]; [Dentler et al. 2024]; [Jaillant 2024]; [Jaillant et al. 2025]. For archivists, integration of AI into archival access raises questions about how provenance, authenticity, and privacy will be addressed. The use of AI to mediate discovery and analysis of digital archives must consider the importance of providing researchers with sufficient contextual information about both the collections and the computational processes applied to them to critically evaluate research findings [Colavizza and Jaillant 2026]. At the same time, these studies reveal persistent barriers. Across projects, Jaillant and colleagues stress that the “lack of collaboration between archivists, librarians and AI developers is one of the most significant impediments to the ethical and inclusive application of AI in collections — especially in collections that contain sensitive historical materials” [Jaillant et al. 2025, Conclusion, ¶2]. To address this, they recommend interdisciplinary training for archivists, education for AI developers on archival requirements, and professional guidelines to ensure AI supports diversity and inclusivity in archival holdings [Jaillant et al. 2025].
Taken together, humanities-led projects demonstrate the range of possibilities for AI integration to enhance access and use of digital collections. They show how AI can be used for discovery and analysis in novel ways, but also how important it is to approach using frameworks of collaboration, ethics, and literacy. A key takeaway is the importance of user-centered design and interdisciplinary partnership.

The I Trust AI Initiative

The I Trust AI Project (2021-2026) is the most recent phase of the International Research on Permanent Authentic Records in Electronic Systems (InterPARES), a series of international, multidisciplinary research projects focused on the creation and preservation of trustworthy digital records. Building on research initiated by The Preservation of the Integrity of Electronic Records (1994-1997) [2] , InterPARES launched in 1999 by investigating the requirements for preserving authentic records created and maintained in electronic systems. Subsequent phases InterPARES 2, 3 & 4 expanded the research to complex digital records and different contexts of creation and preservation, and later to records created and managed in networked and online environments. Across these phases, InterPARES developed theoretical and methodological knowledge, guidelines, and requirements to assist records creators and preservers in establishing and maintaining the trustworthiness of digital records over time. [3]
The current phase, I Trust AI is a multi-national, interdisciplinary effort to design and leverage AI in support of the ongoing availability and accessibility of trustworthy public records. [4] The project is co-directed by Luciana Duranti, Professor Emerita at the University of British Columbia and Muhammad Abdul-Mageed, Canada Research Chair in Natural Language Processing and Machine Learning. With 87 partner organizations across four continents and 257 researchers from universities, government agencies, archival institutions, and industry, the project demonstrates cross-domain collaboration aimed at building capacity in archival education and professional practice [I Trust AI 2024b]. The international scope of I Trust AI includes participation by researchers and partner organizations from North America, Europe, Latin America, Africa, China, and the Middle East. Research conducted outside North America and Europe includes studies of AI and recordkeeping in Brazil [Berrío-Zapata et al. 2022] and Mexico [Barnard Amozurrutia and Enríquez Rodríguez 2023], AI and appraisal in Latin America [Barnard Amozurrutia and Enríquez Rodríguez 2023], AI in records and archives management in China [Xie 2022], and advancing recordkeeping professionals in Africa [Katuu 2024]. [5] This geographic diversity contributes perspectives from different legal, cultural, and technological contexts to the project’s exploration of responsible AI for records and archives.
Within the InterPARES framework, trustworthiness has a specific meaning. A trustworthy digital record is simultaneously reliable, accurate, and authentic: it can stand for the facts to which it attests, its content is correct and free from error or distortion, and it is what it purports to be and has not been corrupted or tampered with [Bushey 2024a]. These qualities depend on controls over records creation, transmission, management, and preservation, making trustworthiness an ongoing recordkeeping concern rather than a determination made only when records are accessed and used.
I Trust AI knowledge mobilization occurs through in-person and virtual plenaries, workshops on ML fundamentals tailored to archivists, and discussions that translate archival principles into computational requirements. A major effort has been directed towards explaining provenance, trustworthiness, and privacy requirements to AI developers so that future systems reinforce rather than compromise archival principles. This has proven more challenging than anticipated due to the theoretical differences between archival science and computer science and the steep technical learning curve archivists must navigate to critically assess AI systems and tools and participate in their design or application [Bushey 2024b].
The project is organized into archival function-based working groups, with early studies exploring expectations of AI and AI use in recordkeeping contexts. Findings have already contributed to the creation of an emerging AI and archives community of practice [Bushey et al. 2025], the development of an AI literacy curriculum for records managers and archivists [Fewster and Arias-Hernández 2024a]; [Arias-Hernández and Rockembach 2025] and a prototype for a secure, privacy-protected, AI-enabled online platform for accessing digital records across institutions [Lemieux et al. 2025]. These outcomes highlight three persistent challenges for archives: safeguarding trustworthy records; aligning technological adoption with ethical frameworks that resist extractive processes; and overcoming the steep AI knowledge and skills barrier within the archival profession.

Survey on Digitization and AI

In spring 2023, I Trust AI researchers in the working group on reference and access conducted a mixed‑methods survey on digitization and AI to gather current data on the state of digitization projects in archives and the use of AI-based tools by archivists and vendors [Bushey et al. 2025]. The survey is part of a larger study exploring potential low-cost AI solutions for digitization, recognizing that many organizations face cost barriers when conducting digitization projects. Led by Eng Sengsavang, Reference Archivist at UNESCO, the study is guided by ethical approaches to the uses and development of AI, as presented in the UNESCO Recommendation on the Ethics of AI (2022), which outlines ethical principles for AI development and use. The following discussion highlights key findings and situates them within the UNESCO framework of values and ethical principles [Bushey et al. 2025, 18—23].
A total of 127 responses were received, with the majority of participants based in North America and Europe. Respondents represented professionals engaged in creating and managing digitized archives and documentary heritage materials, with archivists forming the largest group (31.3%), followed by librarians (17.6%) and a combined group of records managers and digitization/ collections management staff (24.8%). Most worked in governmental or public organizations (59.4%), with smaller proportions in the private sector (21.7%), non-governmental organizations (7.4%), and international or intergovernmental organizations (5.1%).
When asked about digitization policies and legal and regulatory requirements specific to digitization within their organizations, just over half (53.5%) reported having a formal policy, and slightly fewer (44.2%) were aware of specific legal and regulatory requirements, primarily copyright, privacy, and data protection. Significantly more respondents relied on technical and internal guidelines for digitization, such as ISO/TR 13028:2010 and FADGI. These findings highlight the importance of privacy and data protection in the selection of records for digitization and dissemination online. Looking ahead, UNESCO recommends privacy impact assessments of AI systems and tools to ensure that personal information is protected and processed in accordance with data protection principles and standards, including informed consent [UNESCO 2022, 21]. Yet survey respondents using AI in digitization (35%) cited recommendations and ease of use as their main criteria; none reported conducting privacy impact assessments.
Most respondents (75%) agreed that digitization is important for preservation and access. They attributed digitization projects to increased global access (31.5%) and higher use within their organizations (11.5%). Still, nearly two-thirds of respondents (63.6%) reported that their organizations had digitized less than one-quarter of analog materials in their department or unit. Respondents identified funding shortages (21.8%), staffing challenges (16.1%), and limited expertise and training as major barriers. Half emphasized the importance of IT staff in supporting archivists during digitization, suggesting that technical collaboration is critical for project success.
Finding on AI use revealed uneven adoption. Nearly half of respondents work in organizations that have yet to consider AI tools. Of the thirty-five percent that are using AI, half relied on commercial products and half on custom-built tools or systems. The most common applications were OCR/HTR (17.8%), speech-to-text transcription (12.1%), and CV (6.5%). Survey respondents were not always able to identify the AI tool being used for digitization activities, yet ChatGPT, Tesseract, and Transkribus were mentioned with the greatest frequency. Reported benefits included improved efficiency, enhanced searchability, and increased access. At the same time, respondents identified negative impacts of AI tools on digitization, including inaccuracies, poor OCR quality, and lengthy implementation processes. Notably, the survey did not ask whether end users of digitized archives were informed that AI tools had been applied, which is a gap UNESCO identifies as critical for transparency and accountability and contributing to the trustworthiness of records [UNESCO 2022].
Responses to questions about risks and barriers to use reflect both the nascency and complexity of AI in archival contexts. Where risks were identified, concerns centered on bias and ethical implications. Lack of staff training, time, and funding were significant barriers. Yet optimism remained as respondents praised AI’s speed in completing repetitive tasks and improving item-level access, while cautioning that human oversight and quality control were essential. These findings align with UNESCO’s call for human oversight of AI systems, limited delegation of control, and greater AI literacy for the public (pp.22-23).
A report of the findings was shared with respondents and the wider archival and cultural heritage community [Bushey et al. 2025]. The positive reception, including requests for presentations of the findings, underscores the importance of this work. As part of the larger study, the research team is exploring capacity-building through the creation of a community of practice. This approach would allow professionals to share AI tools and use cases while collectively addressing concerns about bias and the ethical implications of applying AI to digitization, particularly for records created by or about Indigenous Peoples and marginalized groups. Such an initiative would align with UNESCO’s principle of “multi-stakeholder and adaptive governance collaboration” [Bushey et al. 2025, 23] and point to a path forward where AI adoption is both practical and ethically accountable.

AI Literacy Framework and Competencies

As discussed throughout this article, findings from both humanities- and archives-led initiatives highlight a persistent gap in AI knowledge and skills, underscoring the need for archival professionals to develop AI competencies (knowledge, skills, and attitudes) to support the integration of AI into archival workflows. In early 2021, an I Trust AI study, led by Richard Arias-Hernández, Associate Professor of Teaching at the University of British Columbia, undertook a literature review to identify use cases, records professionals already working with AI, and potential teaching materials to support competency development [Arias Hernández, Fewster, and Penniman 2024]. Following a series of conference presentations that included audience surveys and real-time feedback, an AI literacy framework was established [Arias-Hernández and Rockembach 2025]. The framework identifies six areas with related competencies: (1) general AI concepts and theories (e.g., ML, NLP, CV); (2) convergence of AI literacy with archives/records‑management literacy; (3) literacy for digitization and datafication with AI; (4) computing literacy related to ethics (e.g., bias, transparency); (5) user‑centric AI literacy; and (6) organizational AI literacy, including collaboration, adoption, and governance. Importantly, the framework emphasizes that AI/ML competencies must be ethical, responsible, transparent and explainable ([Arias-Hernández and Rockembach 2025, 22], requiring “concerted efforts in education, management engagement, resources allocation, technology development, and system integration” [Arias-Hernández and Rockembach 2025, 21].
In late 2024, the first Teachable AI for the Archival Professions module was released [Fewster and Arias-Hernández 2024a]. Reflecting the goals of I Trust AI, the series is designed to train archivists and students to effectively leverage AI in their work. Eight modules have been discussed, each based on case studies involving the application of AI to archives (e.g., text, audio, images) within records management or archival repository contexts. The topics are mapped to a competencies framework. As of writing, five modules are available in draft under a CC BY-NCSA 4.0 license on the I Trust AI website [I Trust AI 2024b]. Module 1, Introduction to AI for Archival Professionals; Module 2, Critical AI/ML, Indigenous AI/ML, and AI/ML Ethics for Archival Professionals; Module 3, AI/ML for Processing Textual Records in Archives; Module 4, AI/ML for Processing Image-based Records in Archives; and Module 5, AI/ML for Processing Audiovisual Records in Archives.
Each module is structured with learning objectives, required readings, activities, and assignments that can be integrated into archival education and professional training. The timeliness of this curriculum is reinforced by the rapid uptake of AI in higher education. For example, the recent partnership between California State University (CSU) and OpenAI has provided unlimited access to ChatGPT Edu for 500,000 students and faculty across 23 campuses [OpenAI 2025]. This initiative sparked a frenzy of optional webinars and training sessions to increase AI literacy for educators. For CSU faculty in the School of Information at San José State University, where archival studies are taught, the Teachable AI modules are not only timely and valuable but critical for situating AI literacy within a robust ethical framework.

Clio-X: Privacy Preserving AI Access

Records professionals have long balanced providing access to archival holdings with protecting the privacy of creators and subjects in line with donor agreements, legal, and regulatory requirements, and more recently cultural protocols. Digitization and datafication have multiplied the volume of records potentially containing personally identifiable information (PII), raising new challenges. While digital archives can be accessed and analyzed as big data, the extent of that access must comply with institutional, national, regional and international frameworks for privacy and data protection.
In early 2023, I Trust AI researchers led by Victoria Lemieux, Professor of Archival Science at the University of British Columbia, founder of Blockchain@UBC, and co-founder of Computational Archival Science (CAS), launched a study examining privacy in digital records containing PII and the potential impact of using AI tools and techniques to address it [Lemieux et al. 2024]. Their findings suggested that privacy enhancing technologies (PET), an emerging class of technologies designed to default toward protecting sensitive information in archival holdings, could be combined with visualization tools to “allow for AI-enabled analysis of archival documents without requiring that researchers gain direct access to the records” [Lemieux et al. 2024, 3]. Such techniques hold promise for balancing privacy with accessibility in AI-enabled archival discovery and analysis [Lemieux et al. 2024, 30].
The next phase of this research launched in early 2025, produced the Clio-X prototype [Lemieux et al. 2025]. Clio-X, is a decentralized Web3 digital solution for privacy-preserving AI-enabled archival reference and access. It is designed to address the growing burden of sensitivity reviews required to redact or anonymize PII before release of records to researchers. To date, many AI applications intended to assist with data protection have met with mixed success, often requiring extensive human oversight and, in some cases, risking data leaks if models are hacked and training data exposed [Lemieux et al. 2025, 3]. By contrast, Clio-X integrates privacy-preserving machine learning with blockchain-supported governance and data visualization tools. This architecture allows archivists to share copies of their records within a trusted environment, offering researchers access to usable data while preventing direct exposure of sensitive information [Lemieux et al. 2025, 7].
User testing of the Clio-X prototype revealed several priorities: an intuitive interface accessible to non-technical users, explainable AI (XAI) features providing transparency about algorithms used in the platform, and training resources to support user adoption [Lemieux et al. 2025, 20]. Feedback also emphasized the importance of clear communication about how the system protects privacy and security, alongside the need to establish user communities for mutual support and collaboration [Lemieux et al. 2025, 21]. As discussed earlier, involving communities and interested parties in system design fosters inclusion and equitable benefit-sharing. Lemieux et al. conclude that “ongoing transparency, strong collective governance and social proof from credible institutions...” [Lemieux et al. 2025, 22] will be essential for building trust and encouraging widespread adoption.
Clio-X represents not only an innovative technical prototype but a new business model for accessing digital archives, one that resists ceding governance and technological control to data brokers and corporate platforms. By enabling secure, decentralized, and community-governed access, Clio-X points to a future where archivists can provide granular levels of access to vast digital holdings, along with computational tools to support leading edge research, while safeguarding privacy, transparency, and trust.

IV. Conclusion

The integration of AI into archival functions offers the promise of enhanced access and use of volumes of digitized and born digital holdings. Yet, as this article has argued, three challenges remain central: protecting the trustworthiness of records, engaging with critical ethical frameworks to resist extractive practices, and overcoming the steep AI literacy barrier within the archival profession.
The initiatives discussed in this article demonstrate the breadth of exploration. Humanities-led projects have tested AI tools for discovery and analysis, highlighting the importance of anticipating user experience and addressing bias and transparency early in system design. The I Trust AI project (2021–2026) demonstrates interdisciplinary collaboration, bringing together archival scholars, AI experts, legal specialists, and records management and archival professionals to align tool development with archival principles of provenance, context, and rights management. The AI literacy and competencies framework [Arias-Hernández and Rockembach 2025] responds directly to the need for interdisciplinary training, ensuring archivists can critically evaluate AI systems and participate in their selection, design, or integration. The survey on digitization [Bushey et al. 2025] offers a snapshot of institutional readiness, documenting barriers such as limited training, funding, and infrastructure alongside the potential for AI to assist in labor-intensive workflows, such as transcription, metadata extraction, and image processing. The Clio-X prototype [Lemieux et al. 2025] pushes the field toward privacy-preserving, explainable, secure, decentralized community-governed AI access platforms for multi-institutional environments. Ultimately, archivists are stewards responsible for ensuring that AI adoption strengthens—not compromises—access to trustworthy records; and that relationships between community members, interested parties, rights holders, and affected groups are placed at the center. If AI systems and tools are integrated into archival activities, they should improve archival work and support archivists in balancing privacy and rights (including data sovereignty) while enhancing access and use. The risks associated with AI are a reminder of the limitations of existing technology and a provocation to reimagine archival access platforms that integrate computational tools for the benefit of archivists and humanities researchers. Digital humanists have an important role in these collaborations as both users of digital archives and partners in the design and evaluation of AI-enabled access systems. Their experience applying computational methods to digital archives can help identify new approaches to discovery and analysis and reveal where technological systems obscure context or constrain interpretation. As [Colavizza and Jaillant 2026] demonstrate in their AI readiness guidelines for archival contexts, collaboration among archivists, digital humanists, and computer scientists can bring together knowledge of records and their contexts, research practices and user needs, and technical expertise in the development of AI systems. The future of interdisciplinary collaboration between archivists, AI developers, and humanities scholars has arrived.

Acknowledgements

The author wishes to thank the anonymous peer reviewers for their thoughtful and constructive feedback, which helped strengthen the clarity and focus of this article.

Notes

[1]  Archives and Manuscripts 47(3), 2019; Archives, Access and Artificial Intelligence, 2022; Special Issue on Computational Archival Science, Archival Science (1) & (3), 2022; AI & Society, 2022; Journal on Computing and Cultural Heritage 4, 2023; Digital Humanities Quarterly 18(2), 2024; AI & Society, 2025.
[2]  Also known as The UBC Project and The UBC-MAS Project. A concise discussion of the project goals, methodological framework, conceptual analysis of records, reliability and authenticity, and major findings can be found in: Luciana Duranti and Heather MacNeil, 1996, The Protection of the Integrity of Electronic Records: An overview of the UBC-MAS Research Project, Archivaria,42, 46-67. The research project was undertaken by researchers at the University of British Columbia (UBC) in collaboration with the United States Department of Defense. One of its products was the DoD Standard 5015.2 Design Criteria Standard for Electronic Records Management Software Applications.
[3]  Research products and dissemination for all four phases of InterPARES can be accessed through the main InterPARES Project website: https://www.interpares.org/.
[4]  The current phase, I Trust AI, has a different website: https://interparestrustai.org/
[5]  I Trust AI research products and dissemination are ongoing. For an up-to-date list, see: https://interparestrustai.org/trust/research_dissemination

Works Cited

Crossing Fonds (n.d.) The crossing fonds digital archives symposium. Available at: https://crossingfonds.com/symposium-2024/. Accessed: 13 August 2025.
Whitelaw, M. (2015) “Generous interfaces for digital cultural collections”, Digital Humanities Quarterly, v.9(1). Available at: https://dhq.digitalhumanities.org/vol/9/1/000205/000205.html.
AEOLIAN Network (n.d.) AEOLIAN Network (n.d.) “Home”. Available at: https://www.aeolian-network.net/. Accessed: 13 August 2025.
AURA Network 'Home' (n.d.) AURA Network (n.d.) “Home”. Available at: https://www.aura-network.net/. Accessed: 13 August 2025.
AURA Network (n.d.) AURA Network (n.d.) “Events”. Available at: https://www.aura-network.net/events/. Accessed: 13 August 2025.
Acker and Kriesberg 2020 Acker, A. and Kriesberg, A. (2020) “Social media data archives in an API-driven world”, Arch Sci 20(305). Available at: https://doi.org/10.1007/s10502-020-09343-y.
Angelova et al. 2021 Angelova L., Ogden B., Craig J., Chandrapal H., and Manandhar D. (2021) “Deep discoveries: A towards a national collection foundation project final report”, Zenodo. Available at: https://doi.org/10.5281/zenodo.5710412.
Arias Hernández, Fewster, and Penniman 2024 Arias Hernández, R., Fewster, K. and Penniman, S. (2024) “Artificial intelligence and machine learning competencies for the archival professions”, Proceedings of the Association for Information Science and Technology, 61, pp. 36–43. Available at: https://doi.org/10.1002/pra2.1006. Accessed: 13 August 2025.
Arias-Hernández and Rockembach 2025 Arias-Hernández, R. and Rockembach, M. (2025) “Building trustworthy AI solutions: integrating artificial intelligence literacy into records management and archival systems”, AI & Society. Available at: https://doi.org/10.1007/s00146-025-02194-0. Accessed: 13 August 2025.
Barnard Amozurrutia and Enríquez Rodríguez 2023 Barnard Amozurrutia, A., and Enríquez Rodríguez, L. L. (eEds.). (2023) Perspectivas y novedades en el ámbito de la archivística. Editorial Porrúa; INFOCDMX.
Barnard et al. 2023 Barnard, A., Mendoza Luz, A., Fenoglio, N., and Umaña, R. (14 October 2023) “Report on Appraisal and Disposition y Latin America”, ITrustAI. Available at: https://interparestrustai.org/assets/public/dissemination/AA01-SG05AppraisalinLatinAmericafinaloct112023.pdf. Accessed: 3 September 2026.
Bastian and Flinn 2020 Bastian, J.A., and Flinn, A. (eds.) (2020) Community archives, community spaces: Heritage, memory and identity. Facet Publishing, UK.
Benoit and Eveleigh 2019 Benoit, E., and Eveleigh, A. Eds. (2019) Participatory archives: Theory and practice. Facet Publishing, UK.
Berrío-Zapata et al. 2022 Berrío-Zapata, C. et al. (2022) “The Arquive of Tatuoca Magnetic Observatory Brazil: from paper to intelligent bytes”, 2022 IEEE International Conference on Big Data (Big Data), Osaka, Japan, pp. 2482-2489, Available at: https://www.semanticscholar.org/paper/The-Arquive-of-Tatuoca-Magnetic-Observatory-Brazil%3A-Berr%C3%ADo-Zapata-Silva/912e6fa9898a3a731e73210c16efc7d60696313d.
Bushey 2023 Bushey, J. (2023) “AI-Generated Images as an Emergent Record Format”, IEEE Big Data 2023 Proceedings, pp. 2020–2031.
Bushey 2024a Bushey, J. (2024a) “Digital images: Trustworthiness at the dawn of ai”, in Rogers, C. and Wieland, A. (eds) Archival science in interdisciplinary theory and practice. Lanham: Rowman & Littlefield, pp. 113–128.
Bushey 2024b Bushey, J. (2024b) “Envisioning archival images with artificial intelligence”, Archeion no.125, pp. 33-54. Available at: https://doi.org/10.4467/26581264ARC.24.007.20202.
Bushey et al. 2025 Bushey, J., De Souza, M., Fukushima, K., Igésias, D., Riess, M., Sengsavang, E., Stancic, H., and Trbusic, Z. (2025) “Report on the survey “digitization and artificial intelligence for archives and documentary heritage materials"”. Final. InterPARES Trust AI. Available at: https://interparestrustai.org/assets/public/dissemination/RA03-InterPARESAI-Survey_Report_FINAL.pdf. Accessed: 13 August 2025.
Carroll et al. 2020 Carroll, S., Garba, I., Figueroa-Rodríguez, O., Holbrook, J., Lovett, R., Materechera, S., Parsons, M., Raseroka, K., Rodriguez-Lonebear, D., Rowe, R., Sara, R., Walker, J., Anderson, J. and Hudson, M. (2020) “The CARE principles for Indigenous data governance”, Data Science Journal, 19, pp. 1–12. Available at: https://doi.org/10.5334/dsj2020-042. Accessed: 13 August 2025.
Caswell 2024 Caswell, M. (2014) “Toward a survivor-centered approach to records documenting human rights abuse: lessons from community archives”, in Special Double Issue on Archives and Human Rights, Caswell, M. (ed.) Archival Science, v.14 (3-4). P. 307-322.
Caswell and Cifor 2016 Caswell, M. and Cifor, M. (2016) “From human rights to feminist ethics: Radical empathy in the archives”, Archivaria, 81, pp. 23–43.
Caswell and Cifor 2019 Caswell, M. and Cifor, M. (2019) “Neither a beginning nor an end: Applying an ethics of care to digital archival collections”, in Lewi, H., Smith, W., vom Lehn, D. and Cooke, S. (eds) The Routledge international handbook of new digital practices in galleries, libraries, archives, museums and heritage sites. London: Routledge, pp. 244–258.
Caswell and Cifor 2021 Caswell, M. and Cifor, M. (2021) “Revisiting a feminist ethics of care in archives: An introductory note”, Journal of Critical Library and Information Studies, 3(2). Available at: https://doi.org/10.24242/jclis.v3i2.162.
Caswell, Cifor, and Ramirez 2016 Caswell, M., Cifor, M., and Ramirez, M. (2016) “"To suddenly discover yourself existing”: Uncovering the impact of community archives”, The American Archivist, 79(1), 56-81. Available at: http://dx.doi.org/10.17723/0360-9081.79.1.56.
Christen 2019 Christen, K. (2019) ““The songline is alive in Mukurtu”: Return, reuse, and respect”, Archival Returns: central Australia and Beyond. LD&C Special Publication 18, University of Hawai’i Press.
Colavizza and Jaillant 2026 Colavizza, G., and Jaillant, L. (2026) “AI ready: how to prepare archives for Artificial Intelligence, improve discoverability, and enhance access”, Archives and Records, 1–22. Available at: https://doi.org/10.1080/23257962.2026.2675937.
Colavizza et al. 2021 Colavizza, G., Blanke, T., Jeurgens, C., and Noordegraf, J. (2021) “Archives and AI: An overview of current debates and future perspectives”, Journal on Computing and Cultural Heritage, 15(1), article 4., pp. 1-15. Available at: https://dl.acm.org/doi/10.1145/3479010.
Cook 2013 Cook, T. (2013) “Evidence, memory, identity, and community: four shifting archival paradigms”, Archival Science, 13, pp. 95–120. Available at: https://doi.org/10.1007/s10502-012-9180-7. Accessed: 13 August 2025.
Dentler et al. 2024 Dentler, J., Jaillant, L., Foliard, D. and Schuh, J. (2024) “Sensitivity and access: Unlocking the colonial visual archive with machine learning”, Digital Humanities Quarterly, 18(3). Available at: https://www.digitalhumanities.org/dhq/vol/18/2/000742/000742.html.
Dryden 2014 Dryden, J. (2014) “The role of copyright in selection for digitization”, The American Archivist, 77(1), pp. 64-95.
Dryden 2017 Dryden, J. (2017) “The meaning of publication in Canadian copyright law: An archival perspective”, Archivaria, 83, pp. 1–34. Available at: https://archivaria.ca/index.php/archivaria/article/view/13598. Accessed: 13 August 2025.
Du, Le, and Honig 2024 Du, L., Le, B., and Honig, E. (2024) “Probing historical image contexts: Enhancing visual archive retrieval through computer vision”, Journal of Computing & Cultural Heritage, 16(4), Article 84. Available at: https://doi.org/10.1145/3631129.
Duff et al. 2013 Duff, W., Flinn, A., Suurtamm, K.E., and Wallace, D.A. (2013) “Social justice impact of archives: A preliminary investigation”, Archival Science v.13(4), p.317-348.
FADGI 2023 FADGI (Federal Agencies Digital Guidelines Initiative) (2023) Technical guidelines for digitizing cultural heritage materials. Third Edition. Available at: https://www.digitizationguidelines.gov/guidelines/FADGI%20Technical%20Guidelines%20for%20Digitizing%20Cultural%20Heritage%20Materials_3rd%20Edition_05092023.pdf. Accessed: 13 August 2025.
Feliciati 2023 Feliciati, P. (2023) “Archival users and AI tools for reference and access: a study within the InterPARES Trust AI project”, Italian Journal of Library Science, Archival Science and Information Science, 14(3). [In Italian]. Available at: https://doi.org/10.36253/jlis.it-558. Accessed: 13 August 2025.
Fewster and Arias-Hernández 2024a Fewster, K. and Arias-Hernández, R. (2024) “Teachable ai for the archival Pprofessions – Module 1: Introduction to artificial intelligence for the archival professions”, I Trust AI. Available at: https://interparestrustai.org/assets/public/dissemination/AD_01_Module1_v_1_2_202410291.pdf. Accessed: 13 August 2025.
Fewster and Arias-Hernández 2024b Fewster, K. and Arias-Hernández, R. (2024) “Teachable AI for the archival professions – Module 2: Critical AI/ML, Indigenous AI/ML, and AI/ML ethics for archival professionals”, I Trust AI. Available at: https://interparestrustai.org/assets/public/dissemination/AD_01_Module2_v_1_0_20241211.pdf. Accessed: 2 September 2025.
Fewster and Arias-Hernández 2025a Fewster, K. and Arias-Hernández, R. (2025) “Teachable ai for the archival professions – AI/ML for processing textual records in archives”, I Trust AI. Available at: https://interparestrustai.org/assets/public/dissemination/AD_01_Module3_v_1_0_202504141.pdf. Accessed: 2 September 2025.
Fewster and Arias-Hernández 2025b Fewster, K. and Arias-Hernández, R. (2025) “Teachable ai for the archival professions – AI/ML for image-based records in archives”, I Trust AI. Available at: https://interparestrustai.org/assets/public/dissemination/AD_01_Module4_v_1_0_20240918.pdf. Accessed: 8 September 2026.
Fewster and Arias-Hernández 2026 Fewster, K. and Arias-Hernández, R. (2026) “Teachable ai for the archival professions – AI/ML for processing audiovisual records in archives”, I Trust AI. Available at: https://interparestrustai.org/assets/public/dissemination/AD_01_Module5_v_1_0_20260327_DRAFT.pdf. Accessed: 8 September 2026.
Friedwald et al. 2024 Friedewald, M., Székely, I. and Karaboga, M. (2024) “Preserving the past, enabling the future: Assessing the European policy on access to archives in the digital age”, Preservation, Digital Technology & Culture, 53(2), pp. 61-71. Available at: https://doi.org/10.1515/pdtc-2024-0003.
GIDA (n.d.) GIDA (Global Indigenous Data Alliance) (n.d.) “CARE principles for Indigenous data governance”. Available at: https://www.gida-global.org/care (Accessed: 13 August 2025).
Giardinetti et al. 2024 Giardinetti, M., Foliard, D., Schuh, J., and Mohamed-Salim, Aissi. (2024) “The EyCon dataset: A visual corpus of early conflict photography”, Journal of Open Humanities Data, 10(1), p. 40. Available at: https://doi.org/10.5334/johd.213. Accessed: 13 August 2025.
Gilliland 2011 Gilliland, A. (2011) “Neutrality, social justice and the obligations of archival education and educators in the twenty-first century”, Archival Science, v.11, pp. 193–209. Available at: https://doi.org/10.1007/s10502-011-9147-0.
Hedges et al. 2022 Hedges, M., Marciano, R., and Goudarouli, E. (2022) “Introduction to the special issue on computational archival science”, Journal on Computing and Cultural Heritage, 15(1), Article 1. Available at: https://doi.org/10.1145/3495004.
Hirtle, Hudson, and Kenyon 2009 Hirtle, P., Hudson, E., and Kenyon, A. (2009) Copyright and cultural institutions: Guidelines for digitization for U.S. libraries, archives and museums. Ithaca, NY: Cornell University Library.
I Trust AI 2024b I Trust AI. (2024b) “Curriculum materials”, in Research dissemination. Available at: https://interparestrustai.org/trust/research_dissemination. Accessed: 2 September 2025.
InterPARES Project (1999-2018) InterPARES Project (1999-2018) Overview. Available at: https://www.interpares.org/. Accessed: 8 September 2026.
InterPARES Trust AI (2024a) InterPARES Trust AI (2024a) Terminology Database. Available at: https://interparestrustai.org/terminology. Accessed: 13 August 2025.
InterPARES Trust AI. (n.d.) InterPARES Trust AI. (n.d.) Home. Available at: https://interparestrustai.org/trust. Accessed: 13 August 2025.
Jaillant 2022a Jaillant, L. (2022a) “More data, less Pprocess: A user-centered approach to email and born-digital archives”, The American Archivist, 85(2), pp. 533-555.
Jaillant 2022b Jaillant, L. (2022b) “Using ai to unlock the potential of digital archives”, Digital Scholarship in the Humanities, 37(2), pp. 571–586.
Jaillant 2022c Jaillant, L. (ed.) (2022c) Archives, access and artificial intelligence: Working with born-digital and digitized archival collections. Verlag: Bielefeld University Press. Available at: https://www.transcript-publishing.com/media/pdf/59/cd/61/oa9783839455845CxlzVP2yr2h76.pdf. Accessed: 25 August 2025.
Jaillant 2023a Jaillant, L. (ed.) (2023a) Special issue: Applying innovative technologies to digitised and born-digital archives, Journal on Computing and Cultural Heritage, 15(4).
Jaillant 2024 Jaillant, L. (ed.) (2024) Special Issue: Using Visual AI applied to Digital Archives, Digital Humanities Quarterly, 18(2).
Jaillant and Rees 2023b Jaillant, L. and Rees, A. (2023b) “Applying ai to digital archives: Trust, collaboration and shared professional ethics”, Digital Scholarship in the Humanities, 38(2), pp. 571–585.
Jaillant et al. 2025 Jaillant, L., Mitchell, O., Ewoh-Opu, E., and Hidalgo Urbaneja, M. (2025) “How can we improve the diversity of archival collections with AI? Opportunities, risks, and solutions”, AI & Society. Available at: https://doi.org/10.1007/s00146-025-02222-z. Accessed: 13 August 2025.
Jules 2016 Jules, B. (2016) “Some thoughts on ethics and DocNow”, Medium. Available at: https://medium.com/documenting-docnow/some-thoughts-on-ethics-and-docnow-d19cfec427f2. Accessed: 28 August 2025.
Jules, Summers, and Mitchell 2018 Jules, B., Summers, E., Mitchell, V., Jr. (2018) “Ethical considerations for archiving social media content generated by contemporary social movements: Challenges, opportunities, and recommendations documenting the now”, White Paper. Available online: https://www.docnow.io/docs/docnow-whitepaper-2018.pdf.
Katuu 2024 Katuu, S. (2024) “Advancing archives and records management professionals in Africa”, Global Knowledge, Memory and Communication, 73(3), pp. 331–353. Available at: https://doi.org/10.1108/GKMC-05-2022-0100.
LUSTRE (n.d.) LUSTRE (n.d.). Outputs. Available at: https://lustre-network.net/outputs. Accessed: 29 July 2025.
Lemieux and Werner 2024 Lemieux, V., and Werner, J. (2024) “Protecting privacy in digital records: The potential of privacy-enhancing technologies”, J. Comput. Cult. Herit. 16(4), Article 83, 18 pages. Available at: https://doi.org/10.1145/3633477.
Lemieux et al. 2024 Lemieux, V., Chung, C., Zhou, C., and Wang, Z. (2024) “Using distant reading to enhance archival access whilst protecting privacy”, Final Report, v.2. InterPARES Trust AI. Available at: https://interparestrustai.org/trust/research_dissemination.
Lemieux et al. 2025 Lemieux, V., Gil, R., Molosiwa, F., Zhou, Q., Li, B., Garcia, R., de la Torre Cubillo, L. and Wang, Z. (2025) “Clio-X: A Web3 solution for privacy-preserving ai acess to digital archives”, Association of Computing Machinery. Accepted for publication.
Mannheimer et al. 2024 Mannheimer S., Rossmann D., Clark J., Shorish Y., Bond N., Scates Kettler H., Sheehey B., Young S.W.H. (eds.) (2024) Special Issue: Responsible AI in Libraries and Archives, Journal of eScience Librarianship, 13(1).
Marciana 2022 Marciana, R. (2022) “Towards a new discipline of computational archival science”, in Jaillant, L. (ed.) Archives, access and artificial intelligence. Bielefeld: Bielefeld University Press, pp. 205–216.
Marciano et al. 2018 Marciano, R., Lemieux, V., Hedges, M., Esteva, M., Underwood, W., Kurtz, M. and Conrad, M. (2018) “Archival records and training in the age of big data”, in Percell, J., Sarin, L.C., Jaeger, P.T. and Bertot, J.C. (eds) Re-envisioning the MLS: Perspectives on the future of library and information science education. Advances in Librarianship, 44B. Emerald Publishing Limited, pp. 179–199. https://doi.org/10.1108/S0065-28302018000044B010.
Mordell 2019 Mordell, D. (2019) “Critical questions for archives as (big) data”, Archivaria, 87, pp. 140–161. Available at: https://archivaria.ca/index.php/archivaria/article/view/13673. Accessed: 13 August 2025.
Moss, Thomas, and Gollins 2018 Moss, M., Thomas, D. and Gollins, T. (2018) “The reconfiguration of the archive as data to be mined”, Archivaria, 86, pp. 118–151. Available at: https://archivaria.ca/index.php/archivaria/article/view/13646.
National Centre for Truth and Reconciliation (n.d.) National Centre for Truth and Reconciliation (NCTR) (n.d.) About the Archives. Available at: https://nctr.ca/records/view-your-records/archives/. Accessed 29 August 2025.
OpenAI 2025 OpenAI (2025) “OpenAI and the California State University system bring ai to 500,000 students and faculty”, Open AI . Available at: https://openai.com/index/openai-and-the-csu-system/.
Owens 2018 Owens, T. (2018) The theory and craft of digital preservation. Baltimore: Johns Hopkins University Press.
Punzalan and Caswell 2016 Punzalan, R., and Caswell, M. (2016) “Critical directions for archival approaches to social justice”, The Library Quarterly, 86(1), pp. 25-42.
Research Data Alliance International Indigenous Data Sovereignty Interest Group 2019 Research Data Alliance International Indigenous Data Sovereignty Interest Group (2019) “CARE principles for Indigenous data governance”, The Global Indigenous Data Alliance. Available at: https://www.rd-alliance.org/wp-content/uploads/2024/03/CARE20Principles20for20Indigenous20Data20Governance_OnePagers_FINAL20Sept2006202019.pdf.
Robinson 2014 Robinson, L. (2024) “Shaking up the archives: Computational archival science accelerates historical research and digital readiness”, College of Information. University of Maryland. Available at: https://ischool.umd.edu/news/shaking-up-the-archives-computational-archival-science-accelerates-historical-research-and-digital-readiness/. Accessed: 31 August 2025.
Rodriguez 2018 Rodriguez, M. (2018) “Mukurtu content management system”, Technical Services Quarterly, 35(4), pp. 406–407. Available at: https://doi.org/10.1080/07317131.2018.1509454. Accessed: 13 August 2025.
Summers 2023 Summers, E. (2023) “Looking forwards”, Medium. Available at: https://medium.com/documenting-docnow/looking-forwards-64cee8436640. Accessed: 28 August 2025.
Towards a National Collection, (A) (n.d.) Towards a National Collection (n.d.,a) About. Available at: https://www.nationalcollection.org.uk/about.
Towards a National Collection, (B) (n.d.) Towards a National Collection (n.d.,b). Deep Discoveries. Available at: https://www.nationalcollection.org.uk/Foundation-Projects.
UK Research and Innovation, (A) (n.d.) UK Research and Innovation (n.d.) “AURA (Archives in the UK/ Republic of Ireland & AI): Bringing together digital humanists, computer scientists & stakeholders to unlock cultural assets”. Gateway to Research. Available at: https://gtr.ukri.org/projects?ref=AH%2FV002341%2F1. Accessed: 30 July 2025.
UK Research and Innovation, (B) (n.d.) UK Research and Innovation (n.d.) AEOLIAN (Artificial intelligence for cultural organisations). Gateway to Research. Available at: https://gtr.ukri.org/projects?ref=AH%2FV009443%2F1. Accessed: 30 July 2025.
UK Research and Innovation, (C) (n.d.) UK Research and Innovation (n.d.) Unlocking our digital past with artificial intelligence (LUSTRE). Gateway to Research. Available at: https://gtr.ukri.org/projects?ref=AH%2FX003132%2F1. Accessed: 29 July 2025.
UNESCO 2022 UNESCO (2022) Recommendation on the ethics of artificial intelligence. Paris, France. Available at: https://unesdoc.unesco.org/ark:/48223/pf0000381137.
United Nations General Assembly 2007 United Nations General Assembly (2007) United Nations declaration on the rights of Indigenous peoples (Resolution A/RES/61/295). United Nations. Available at: https://digitallibrary.un.org/record/606782.
Van Garderen 2009 Van Garderen, P. (2009) “The ICA-AtoM project and technology”, Presentation at the Third Meeting on Archival Information Databases, Association of Brazilian Archivists. Rio de Janiero, Brazil, 16-17 March.
Weinberg 2025 Weinberg, M. (2025) The Glam-E Lab. “Are ai bots knocking cultural heritage offline?”’ v.1.0. Engelberg Center on Innovation Law & Privacy, NYU School of Law. CC BY 4.0.
Whitelaw 2012 Whitelaw, M. (2012) “Towards generous interfaces for archival collections”, Comma, 2. International Conference of Archives, pp. 123-132. Available at: https://mtchl.net/assets/Whitelaw_ICA_GenerousInterfaces.pdf. Accessed: 13 August 2025.
Whitelaw 2019 Whitelaw, M. (2019) “Succession: a generative approach to digital collections”, in Lewi, H., Smith, W., vom Lehn, D. and Cooke, S. (eds) The Routledge international handbook of new digital practices in galleries, libraries, archives, museums and heritage sites. London: Routledge, pp. 389 – 396.
Xie 2022 Xie, S., Siyi L., and Han R. (2022) “Competing with artificial intelligence – can the records and information management profession withstand the challenge?”, Records Management Journal, 32(2), pp. 151–169. Available at: https://doi.org/10.1108/RMJ-08-2021-0033.