The Rising Tide of AI Ethics: A Deep Dive from the Library and Information Science Frontier
The conversation surrounding artificial intelligence has irrevocably shifted. What was once the domain of computer scientists and futurists, focused primarily on capability and efficiency, has now become a mainstream societal debate centered on responsibility and consequence. As AI systems weave themselves into the very fabric of our daily lives—from the algorithms curating our newsfeeds to the autonomous vehicles navigating our streets, and the robots assisting in our hospitals—the ethical questions they raise are no longer hypothetical. They are urgent, complex, and demand answers from a multitude of perspectives. A recent, meticulous analysis published in the Journal of Modern Information has pulled back the curtain on how one crucial, yet often overlooked, discipline is contributing to this global discourse: Library and Information Science (LIS). This research, spearheaded by Kun Huang of Beijing Normal University, Xiaoting Xu of Nanjing University, Anrunze Li, and Feng Xu of the Institute of Scientific and Technical Information of China, reveals that the LIS community is not merely a passive observer but an active, insightful participant, mapping the ethical minefield of AI with a unique blend of humanistic concern and technical understanding.
For decades, the LIS field has been the guardian of information. Its core mission revolves around the organization, management, dissemination, and ethical stewardship of knowledge. Librarians and information scientists have long grappled with issues of privacy, intellectual freedom, equitable access, and the societal impact of information technologies. This deep-rooted expertise in navigating the moral dimensions of the information lifecycle—from creation and collection to storage, use, and disposal—positions the LIS community uniquely to analyze the ethical quandaries posed by AI. AI, after all, is fundamentally an information technology, albeit an immensely powerful and increasingly autonomous one. It thrives on data, operates through algorithms, and manifests as intelligent systems that interact with humans and society. The ethical risks, therefore, are not alien to LIS; they are an amplified, more intricate version of the challenges the field has always faced.
The study by Huang, Xu, Li, and Xu, which analyzed 39 academic papers published between 2015 and 2019 in SSCI-indexed LIS journals, paints a comprehensive picture. It demonstrates that LIS researchers are examining AI ethics with a scope and depth that mirrors the broader, interdisciplinary field. Their work is not confined to a single niche but spans a wide spectrum, investigating ethical issues at three distinct, yet interconnected, levels: data, algorithms, and the AI systems themselves. This multi-layered approach is critical because an ethical failure at any one level can cascade into broader societal harm.
At the data level, the focus is on the raw material that fuels AI. LIS scholars are sounding the alarm on privacy violations that occur during data collection, processing, and sharing. In the sensitive domain of healthcare, for instance, the deployment of care robots that monitor patients 24/7 raises profound questions. These machines collect intimate data on a person’s movements, habits, and even states of undress. Without explicit, informed consent, this constant surveillance is not just a breach of privacy; it is a fundamental assault on human dignity. Similarly, the push for sharing medical data between institutions to improve treatment and research is laudable, but it carries the inherent risk of exposing sensitive health information if data protection measures are inadequate. The research highlights how LIS professionals are attuned to these nuances, understanding that data is not an abstract commodity but a representation of human lives and experiences.
The analysis extends to the political sphere, where the use of citizen data by public sector AI applications can lead to surveillance and the restriction of civil liberties. The potential for data to be manipulated or exploited by bad actors for personal gain is a constant threat. Even in fields like earth science, seemingly benign activities like geographic text analysis can have ethical implications. One study cited in the research points out that standard geographical dictionaries, lacking historical context, can inadvertently erase indigenous place names and perpetuate the legacy of colonialism, causing cultural and social harm. In the home, smart devices collect a treasure trove of personal data, yet users, overwhelmed by lengthy and incomprehensible privacy agreements, often click “agree” without truly understanding the implications, leaving their digital footprints vulnerable. The LIS perspective here is invaluable, as it emphasizes the entire data lifecycle and the ethical responsibilities at each stage, a framework that computer scientists focused solely on model performance might overlook.
Moving up the stack, the algorithm level presents a different set of challenges. Here, the concern is not just about the data itself, but about how it is processed and the decisions that are made. Algorithms, often perceived as neutral and objective, are in fact human creations that can inherit and even amplify societal biases. The LIS research meticulously documents how this plays out in various sectors. In economics, algorithms used to analyze satellite imagery for regional development can lead to discriminatory practices against poorer areas. In employment, AI systems trained on historical hiring data, which often reflects gender imbalances, can perpetuate discrimination by favoring male candidates. In finance, credit-scoring algorithms can exhibit bias based on race, gender, or zip code, creating a vicious cycle that deepens existing inequalities. The research underscores a critical point: algorithmic bias is not a glitch; it is a systemic issue that can have devastating real-world consequences, denying people jobs, loans, and opportunities.
The political arena is again a hotspot, with algorithms used in election campaigns to manipulate voter preferences by altering search results, thereby undermining democratic processes. In the legal system, predictive policing algorithms and tools used by judges to assess recidivism risk have been shown to be biased against minority groups, particularly Black defendants, leading to harsher sentences and perpetuating systemic injustice. In journalism, recommendation algorithms can create “filter bubbles,” subtly guiding users towards content that reinforces their existing beliefs and limiting their exposure to diverse viewpoints, thereby eroding the foundation of an informed citizenry. The LIS contribution here is to frame these issues not just as technical problems to be debugged, but as profound social and ethical failures that require societal solutions.
The most complex and perhaps most discussed layer is that of the AI system itself—the intelligent agents, robots, and autonomous machines that interact with the world. At this level, the ethical questions become deeply philosophical and existential. The research highlights how LIS scholars are engaging with the profound impact these systems have on human dignity, social relationships, and moral responsibility. In social settings, companion robots designed to provide emotional support can trigger genuine emotional responses in users. For vulnerable populations, such as the elderly or children, this can lead to the formation of one-sided emotional bonds, potentially reducing their desire for real human interaction and causing psychological harm. The question of whether a robot can or should have moral agency is fiercely debated. Should a robot be allowed to disobey a human command? Under what circumstances, if any, should it be permitted to cause harm to a human? And if something goes wrong, who is to blame—the robot, its programmer, its owner, or the corporation that built it? The LIS analysis points out that the current legal and ethical frameworks are ill-equipped to handle these questions, often leading to a dangerous “moral scapegoating” where responsibility is diffused or misplaced, eroding public trust in institutions.
In healthcare, the use of robotic caregivers presents a poignant ethical dilemma. While they can perform physical tasks like feeding or lifting, they cannot provide the human touch, empathy, and compassion that are essential to dignified care. Patients may feel objectified, their sense of self-worth diminished. Studies cited in the research show that elderly individuals under robotic care may feel their autonomy is stripped away, as constant monitoring restricts their freedom of movement. The loss of human contact can lead to profound loneliness and a decline in mental well-being. Some researchers advocate for a “care ethics” approach to robot design, emphasizing relationships, roles, and responsibilities, ensuring that the machine’s purpose is not just efficiency but the holistic well-being of the person it serves.
The military domain presents perhaps the most chilling ethical challenges. The development of autonomous weapons systems that can select and engage targets without direct human intervention forces us to confront the value of human life. Can a machine truly understand the gravity of taking a life? Most ethicists, and the LIS research reflects this consensus, argue that it cannot. The delegation of life-and-death decisions to an algorithm is seen as a profound abdication of moral responsibility. Furthermore, when such systems inevitably cause civilian casualties, the chain of accountability becomes impossibly tangled, making justice and retribution nearly unattainable. In the realm of transportation, self-driving cars face the infamous “trolley problem” in real-time, forcing them to make split-second, morally fraught decisions during accidents. The question of who is liable—the passenger, the manufacturer, the software developer—remains legally murky and ethically fraught. The LIS perspective here is crucial in reminding us that technology should serve humanity, not the other way around, and that the pursuit of technological advancement must never come at the cost of our core human values.
Faced with this daunting landscape of ethical risks, the LIS community is not content with merely identifying problems; it is actively proposing solutions. The research categorizes these solutions into two broad, complementary approaches: social methods and technical methods. Social methods focus on establishing the ethical guardrails within which AI must operate. This involves embedding moral principles like fairness, justice, and respect for human rights into the very design and deployment of AI systems. It means advocating for robust legal and regulatory frameworks, such as adapting international treaties on data protection and algorithmic transparency to the AI context, or ensuring that autonomous military systems comply with the laws of war. It also involves fostering a culture of ethical responsibility within organizations that develop and deploy AI, creating clear guidelines and accountability structures.
Technical methods, on the other hand, seek to build ethics into the technology itself. This is a more hands-on, engineering-focused approach. At the data level, it involves developing better techniques for anonymizing sensitive information, creating standards for what data can be collected and how it can be used, and employing technologies like blockchain to enable secure, privacy-preserving data sharing between institutions. At the algorithmic level, it means designing “unsupervised learning” models that can detect and flag potential biases in datasets before they are used for training, or creating “moral decision-making models” that can guide an AI system’s behavior in ethically complex situations. At the system level, researchers are exploring architectural solutions, such as building an “ethical core” into the AI’s operating system—a dedicated module designed to simulate ethical scenarios and prevent harmful actions. Some even propose endowing robots with a form of “moral competence,” programming them to monitor moral language, evaluate events against ethical rules, and engage in moral reasoning.
What makes the LIS contribution so vital is its dual perspective, seamlessly blending humanism with technicism. On one hand, it adopts a human-centered view, prioritizing the impact of AI on individual rights—privacy, autonomy, dignity, freedom—and on the social fabric, including cultural heritage and interpersonal relationships. On the other hand, it engages with the technology on its own terms, understanding the mechanics of data, the logic of algorithms, and the architecture of intelligent systems. This allows LIS researchers to speak the language of both ethicists and engineers, acting as crucial translators and mediators in a field that desperately needs interdisciplinary dialogue.
The research by Huang, Xu, Li, and Xu concludes with a forward-looking call to action for the LIS community. While the field has made significant strides, there is much more to be done. They urge their colleagues to deepen their engagement with the full spectrum of ethical issues across all levels and application domains. This means applying the well-established principles of information lifecycle management to the AI context, examining how ethical risks manifest at every stage, from data creation to its eventual deletion. It also means expanding the scope of inquiry into new and emerging areas, such as the ethical implications of AI in social media, mobile applications, and, crucially, within the library and information services themselves. As libraries deploy AI-powered chatbots, robotic book-sorting systems, and intelligent research assistants, they must lead by example, rigorously analyzing the potential impact on their staff, their patrons, and the very mission of the library as a democratic institution.
Furthermore, the authors emphasize the critical need for the LIS field to play a more active role in shaping the ethical norms and guidelines that will govern AI. As governments and international bodies scramble to draft AI ethics charters, the deep, practical, and humanistic insights of the LIS community must be at the table. The goal is not to stifle innovation but to ensure that it is channeled in a direction that benefits all of humanity, upholding our shared values and protecting our most vulnerable. The research serves as a powerful reminder that the development of ethical AI is not a task for technologists alone. It is a collective societal endeavor, and the Library and Information Science discipline, with its unique history and expertise, is poised to be one of its most important architects.
This analysis, published in a leading information science journal, is more than just an academic exercise. It is a clarion call, a detailed map of the ethical terrain, and a testament to the indispensable role that information professionals will play in building a future where artificial intelligence serves as a force for good, guided by wisdom, empathy, and an unwavering commitment to human dignity.
By Kun Huang, School of Government, Beijing Normal University; Xiaoting Xu, School of Information Management, Nanjing University; Anrunze Li, School of Government, Beijing Normal University; Feng Xu, Institute of Scientific and Technical Information of China. Published in the Journal of Modern Information, 2021, Vol. 41, No. 6. DOI: 10.3969/j.issn.1008-0821.2021.06.015