In the age of Big Data and artificial intelligence, internet access is often assumed to be ubiquitous and therefore no longer a pressing concern. Such assumptions, however, overlook the lived realities of many communities for whom internet connectivity remains unreliable, expensive, or collectively negotiated. In response to these conditions, the field of minimal computing offers a critical intervention. Minimal computing foregrounds computational parsimony by advocating for the use of the least amount of technical resources necessary to achieve a project’s goals. Within Digital Humanities (DH), this approach has become closely associated with sustainability and accessibility, most notably through practices such as static site generation.

Yet even minimal computing workflows can reproduce structural inequalities when their infrastructures rely on platforms that are inaccessible, ethically compromised, or extractive. For communities that must crowdsource their internet bill, hosting and maintaining web servers locally is often unfeasible. As a result, many DH projects turn to GitHub Pages as a “free” solution for hosting static websites. Since its acquisition by Microsoft in 2018, GitHub has been criticized for its collaboration with U.S. Immigration and Customs Enforcement (ICE), its increasing integration of proprietary AI tools, and its role in consolidating control over scholarly and cultural code infrastructures. These issues are particularly pressing when working with culturally sensitive data, where questions of data colonialism, ownership, long-term access, and digital sovereignty are central.

This paper argues that Codeberg, a non-profit, community-led alternative grounded in free and open-source principles, offers a more ethical and sustainable hosting solution through Codeberg Pages. Despite its strong alignment with DH values, Codeberg remains underutilized in the field. The paper contributes to DH praxis by providing a step-by-step guide for building and hosting static websites on Codeberg Pages using widely adopted tools such as Jekyll. By foregrounding practical implementation, the paper aims to lower technical barriers while encouraging infrastructural choices that reflect ethical commitments.

The methodological discussion is grounded in a cooperative project between the University of Namibia (UNAM) and the Centre for Contemporary and Digital History (C2DH) at the University of Luxembourg. The project seeks to develop an accessible and sustainable digital community archive centered on the social history of the Old Location in Windhoek, Namibia. Until its demolition in 1968, the Old Location was the country’s largest racially segregated Location and a vital cultural and social hub for Namibia’s Black population. The forced relocation of residents to Katutura in 1959 led to protests violently suppressed by the South African colonial government, marking a crucial moment in Namibia’s independence struggle and women’s history. Despite its significance, accessible non-colonial sources on the Old Location remain scarce.

Direct engagement with local communities is therefore central to the project, particularly through the digitization of private archives held by former residents. This data will be made accessible through a digital community archive, hosted on a server at UNAM. To further ensure a sustainable knowledge transfer, we’re offering a six-week-long block seminar for students at UNAM, teaching students the minimal computing practices proposed in this paper.

Scholarly communication and research practices – especially in the humanities – are undergoing profound change. The growing influence of platforms such as Google Scholar, Sci-Hub, and AI-driven entities like OpenAI has transformed access to information, but often at the cost of transparency, source verification, and contextual depth. In this evolving ecosystem, libraries stand out as guarantors of reliability, providing curated, validated resources that remain essential for academic and public knowledge creation.

Digital transformation, open science, and data-driven methods have redefined scholarly work and broadened the remit of libraries far beyond their traditional roles. Today, they serve dual audiences: human readers and computational agents, including large language models (LLMs). Yet libraries face pressing challenges: competition from unverified sources, the ease of turning to generative AI without critical evaluation, and reduced engagement with curated collections. The solution lies in reaffirming their strengths – transparent provenance, methodological rigor, rich metadata, and quality‑controlled access – while integrating innovative digital tools that maintain relevance in times of machine‑assisted research (Horstmann 2022).

The Göttingen State and University Library (SUB Göttingen) and the German National Library (DNB) exemplify this strategic adaptation. As partners in Text+, the German National Research Data Infrastructure (NFDI) consortium for language- and text-based research data, they advance FAIR-compatible infrastructures (Wilkinson et al. 2016). The establishment of a dedicated GND agency within Text+ empowers researchers to contribute directly to authoritative metadata systems (Buddenbohm/Fischer 2023). SUB Göttingen builds on a long tradition of pioneering the digital humanities through infrastructures such as TextGrid and DARIAH-DE, now enhanced with modern computational tools for philological analyses. DNB transforms its vast holdings into derived text formats (Schöch et al. 2020), thus facilitating secure, transparent research on copyrighted materials and exploring their use for training trustworthy LLMs in projects such as CORAL.

Crucially, libraries are reimagining their role in public engagement with science. Through citizen science collaborations, thematic exhibitions, open workshops, and access to datasets, they invite individuals beyond academia to participate in research. This inclusivity bridges the gap between scholarly communities and the wider public, countering the exclusivity of algorithmically filtered knowledge and fostering trust through openness and demonstrable provenance. This talk shows how libraries position themselves as spaces for discussion, co-creation, and informed decision-making, guiding users – whether novice learners or advanced researchers – toward evidence-based understanding rather than unverified digital noise.

The active participation of libraries in 20 of 26 NFDI consortia underscores that they are not passive observers of change but architects of reliable digital knowledge ecosystems (Kurzawe/Stein 2024). By combining expertise in standardization, preservation, and connectivity with innovative strategies for public engagement, libraries offer a coherent answer to the challenges of the digital age. They remain unparalleled in delivering trustworthy information and ensuring that scholarly and societal knowledge prospers amid evolving technologies and shifting communication channels.

This talk explains why working in and accessing libraries is vital for democratic societies, especially in today’s accelerating information flows. Libraries stand as lighthouses guiding scholars and the public toward knowledge rooted in evidence, integrity, and shared understanding.

This paper explores the emergence of "Hybrid Cosmologies" within the digital practices of Siberian Indigenous communities as a strategy for navigating the "shifting grounds" of political and ecological uncertainty. Responding to the conference theme of "infrastructures of engagement," this study examines how digital platforms serve as alternative spaces where marginalized communities co-construct knowledge and assert presence.

The study argues that contemporary Siberian digital creators are engaging in Indigenous Data Sovereignty, effectively operationalizing the CARE principles (Collective Benefit, Authority to Control, Responsibility, Ethics) outside of traditional institutional frameworks. Understood here as the right of Indigenous peoples to govern data about their communities, this practice manifests as reclaiming the power to curate memory archives. However, the paper addresses the paradox of asserting such sovereignty on corporate platforms (YouTube, TikTok), asking how commercial tools can be repurposed into resilient infrastructures for public good.

Drawing on "decolonial computing" (Ali, 2016) and Kathleen Fitzpatrick’s notion of research in conversation with the public, the paper posits that social media functions as a networked medium that challenges static historical narratives. Engaging with Andrew Hoskins’ (2011) "connective turn," the analysis demonstrates how memory becomes a dynamic process of connectivity rather than a stored object. The research employs digital ethnography of content produced by key Siberian influencers, specifically the group Otyken and grassroots eco-activists.

The analysis distinguishes between "performative exoticism" and "deep revitalization." It demonstrates how actors like Otyken strategically blend commercial pop-shamanism with authentic linguistic heritage to "hack" recommendation algorithms. This "digital show" serves as a Trojan horse: capturing global attention to redirect it toward local crises (wildfires, industrial pollution). Ultimately, this paper suggests that these participatory practices are not just forms of resistance, but are actively shaping collective memory and identity, offering a model for how digital humanities can support ethical, community-driven scholarship in authoritarian contexts.

References:
1. Ali, S. M. (2016). A brief introduction to decolonial computing. XRDS: Crossroads, The ACM Magazine for Students, 22(4), 16-21.
2. Hoskins, A. (2011). Media, Memory, Metaphor: Remembering and the Connective Turn. Parallax, 17(4), 19-31.
3. Kukutai, T., & Taylor, J. (Eds.). (2016). Indigenous Data Sovereignty: Toward an agenda. ANU Press.
4. Risam, R. (2018). New Digital Worlds: Postcolonial Digital Humanities in Theory, Praxis, and Pedagogy. Northwestern University Press.

The transition from the textual dominance of the "Gutenberg Galaxy" to an ocularcentric era characterized by synthetic media fundamentally alters the ontology of historical evidence. As generative artificial intelligence becomes capable of fabricating hyper-realistic audiovisual content, the digital public sphere faces a profound epistemic crisis. The boundary between authentic historical records and algorithmic hallucinations is dissolving. While Digital Humanities has traditionally excelled in text-centric hermeneutics, the field currently lacks adequate methodological frameworks to address the multimodal nature of contemporary disinformation. This paper posits that addressing this "blind spot" is not merely a technical challenge but a foundational imperative for the humanities. We argue that the university must reclaim its role as a guarantor of truth by developing sovereign "Infrastructures of Truth."

The current landscape of media verification relies dangerously on "epistemic outsourcing." The advanced tools required to detect synthetic content are almost exclusively proprietary and operated by the same technology corporations that facilitate its creation. This reliance on corporate "black boxes" presents a dual threat to academic integrity: it compromises data sovereignty by processing sensitive research data through opaque commercial APIs and creates dependency on unexplainable algorithms, rendering verification scientifically non-reproducible. If the academy relies on commercial entities to distinguish signal from noise, it surrenders its critical agency.

In response, this contribution proposes the implementation of Local Inference Clusters as a decentralized and university-governed infrastructure for forensic media analysis. Drawing on a case study of detecting "Biological Incongruence" in political discourse, we demonstrate how open-source Large Language Models and multimodal neural networks can be deployed within the secure perimeter of research institutions. This approach ensures strict adherence to ethical standards and GDPR compliance while allowing researchers to inspect the internal reasoning of the models.

Furthermore, this project operationalizes the concept of "hybridity" by dissolving the rigid barrier between computational forensics and humanistic inquiry. We posit that an infrastructure of truth cannot function solely as a technical filter; it must operate as a participatory knowledge space. By transparently visualizing the dissonance between audio signals and facial micro-expressions, we provide a scaffold for public digital literacy. This approach moves beyond binary "true/false" labeling, offering instead a "pedagogy of verification" that empowers citizens to recognize the material traces of synthetic manipulation. Thus, the research infrastructure serves not only to detect falsehoods but to cultivate a more resilient, critically engaged public sphere capable of navigating the synthetic age.

We frame this architecture as a renewal of classical Quellenkritik (source criticism) for the algorithmic age. This methodology, which we term Forensic Hermeneutics, empowers humanists to move beyond the passive interpretation of cultural artifacts to the active verification of their material authenticity. By shifting from external corporate dependencies to resilient internal computing capacities, Digital Humanities can provide a critical public value. This infrastructure of engagement offers independent, transparent, and scientifically rigorous tools to civil society, ensuring that the power to discern reality remains a public good rather than a privatized commodity.