AI, Knowledge Power, and Global Cities: Strategic Implications of Academic Infrastructure Restructuring

When AI Rebuilds the “Operating System” of Knowledge

Artificial intelligence is being written into nearly all infrastructure systems—from traffic control and energy grids to urban safety and public services. Yet, amid this global AI race, a more hidden and even more foundational system is being reassembled: the mechanisms for managing and disseminating scholarly knowledge. Libraries, digital repositories, academic publishing, and even citation indexes—these seemingly neutral, static knowledge infrastructures precisely constitute the underlying framework of global thought production. They determine which research is seen, whose views gain authority, and which knowledge becomes computable assets.

A recent study published in Humanities and Social Sciences Communications is a sober examination of this change. The study avoids the narrative of “how AI creates profits in commercial media” and instead turns to an often-overlooked field: the actual role of artificial intelligence in library information science, digital humanities, and scholarly knowledge dissemination. Using a hybrid method that combines bibliometric analysis with web sentiment analysis, the researchers seek to answer a fundamental question: when AI begins to automatically organize, index, and recommend knowledge, are we prepared to face the reordering it brings?

The Automation of Knowledge and Hidden Power

The study’s key findings are not earth-shattering: AI’s role in scholarly knowledge management mainly appears in three aspects—automated metadata indexing, citation analysis, and intelligent recommendation systems. These aspects all point to an “efficiency imaginary”: using algorithms to replace manual labor in tedious document classification, using models to predict which research readers might find interesting, and using bibliometric tools to evaluate academic influence. This may appear to be mere task automation, but in reality it means that the pathways of knowledge discovery are being redrawn.

In the past, librarians, editors, and citation indexers—as gatekeepers of the knowledge system—had relatively clear, and relatively contestable, judgment criteria. Within an algorithm-driven system, however, classification, ranking, and recommendation have become closed computational black boxes. Researchers may not know why their papers are assigned to a particular category, why they are recommended to certain readers, or why they sit at the margins of the citation network. This hidden power is far harder to challenge than the administrative authority of the past.

The resulting knowledge inequality will be structural. Languages, disciplines, and regions that are well represented in training data will be more easily recognized and recommended by algorithms; researchers from the “Global South,” non-English contexts, and marginal disciplines, however, may face a new kind of “discoverability crisis”—not being rejected, but being silently excluded from the digital horizon.

Why Should Cities Care?

This brings us back to a broader strategic dimension: global urban competition. Knowledge production has never been an abstract activity floating in a void; it gathers in specific urban spaces—universities, research laboratories, digital enterprises, international organizations, and library networks. A city’s position in the global division of labor increasingly depends on whether it can become a hub for knowledge flows, not merely a node for material flows.The penetration of AI technologies into academic infrastructure has brought new governance challenges for research cities. On the one hand, competition among cities such as Boston, San Francisco, Beijing, and London is no longer just a tournament of publication counts or university rankings. City governments need to ask whether their knowledge infrastructure possesses "algorithmic autonomy": in the face of AI tools dominated by commercial giants, are public academic libraries and research institutions capable of developing knowledge management models that are transparent, fair, and aligned with the public interest?

On the other hand, cities are also policy laboratories. From Toronto to Amsterdam, AI ethics frameworks are gradually being implemented at the city level. Yet the vast majority of AI ethics issues still focus on autonomous driving, facial recognition, or automated employment decisions; academic knowledge systems are rarely placed on the public agenda. This absence suggests that urban planners have not yet fully grasped that traditional knowledge spaces such as academic libraries are precisely the "gateways" for a new generation of algorithms——whoever owns this gateway controls the entrance to research and thought.

Mixed Methods: A Mirror for Urban Policy Research

Another contribution of this study lies in the paradigmatic implications of its methodological design. Bibliometric analysis can map the macro trajectories of research systems, but it cannot easily reveal users' real experiences; online text and sentiment analysis, by capturing discussions among scholars, librarians, and the public, adds complex micro-level perceptions to the macro picture. This "system + user" mixed-method approach is, in essence, a human-centered strategy for evaluating infrastructure——it demands that when assessing technological impacts, we do not merely look at metrics, but also listen to voices.

This strategy is equally important for urban and regional planning. Many smart city projects fail precisely because policymakers over-rely on sensor data and efficiency indicators while ignoring residents' perceptions. In the same way, the construction of AI knowledge infrastructure faces a similar danger: automated metadata systems may perform excellently technically, but if scholars and readers lose trust in them, public knowledge systems will rapidly degenerate into the internal spaces of a few institutions.

Accordingly, the study's emphasis on "bias" and "privacy" should not be read as generic ethical rhetoric. It reveals a crucial shift in AI rights: within knowledge infrastructure, privacy is no longer merely a matter of personal data, but also concerns the integrity of collective cognition. When algorithms, based on opaque training data, marginalize certain knowledge lineages, it is not just individual users' preferences that are violated, but the plurality of human knowledge structures itself.

A New Narrative of "Knowledge Sovereignty" Is Needed

The study's calls for interdisciplinary collaboration, greater AI transparency, and solutions to bias and privacy issues may look like routine academic advice. But in the broader context of global technopolitics, we must recognize that these principles are moving from laboratory norms toward national geopolitical strategy. In the AI era, knowledge management is becoming a crossroads of national security and urban competitiveness.For major urban agglomerations in Europe, North America, and East Asia, building reliable public alternatives within digital knowledge systems can lay the groundwork for broad national AI strategies. For the emerging knowledge cities of the Global South in the Middle East, South Asia, Latin America, and Africa, securing a place in the AI-driven knowledge ecology may be more urgent than competing for any single artificial intelligence model. If these cities rely too heavily on external digital platforms to organize their own academic heritage, they will lose not only their "right of discovery" but also control over their knowledge sovereignty.

Conclusion: Long-termism in Knowledge Infrastructure

Common sense misleads us into thinking that libraries, thesis repositories, and journal databases are merely passive content containers during digitization. In fact, they are continuously evolving sociotechnical entities. AI's intervention is not a one-time upgrade of tools but a long-term institutional transformation. It can organize scattered knowledge into organic flows, or it may flatten the ambiguities of knowledge into computational repetition.

The research on which this article is based reminds us that the global academic knowledge system is entering a new "formative period." At this stage, algorithmic systems have relatively strong fault tolerance, and institutional flexibility still exists. If cities and universities around the world now begin to establish systematic AI governance frameworks—incorporating the diversity of knowledge maps, the transparency of algorithmic decision-making, and the Global South's right to participate into their strategic vision—they may avoid, over the next two decades, a landscape in which knowledge infrastructure is monopolized by a single imagination.

Finally, amid the grand narratives of AI transforming cities and realizing the "future city," we should not forget a more fundamental urbanity: the city is a container of memory, learning, and dissent. The academic knowledge infrastructure shaped by artificial intelligence is determining what future cities can remember, what they overlook, and who is allowed to discover. This is no longer a professional topic that belongs only to library and information science scholars; it is a public issue concerning the future of every knowledge society.

Reference: Artificial intelligence in digital media, humanities, and information science: a multidimensional analysis of research trends and user perceptions