𝗪𝗵𝗮𝘁 𝗛𝗮𝗽𝗽𝗲𝗻𝘀 𝗪𝗵𝗲𝗻 𝗘𝘅𝗽𝗲𝗿𝘁𝗶𝘀𝗲 𝗗𝗶𝘀𝗮𝗽𝗽𝗲𝗮𝗿𝘀? 𝗔.𝗜. 𝗮𝗻𝗱 𝘁𝗵𝗲 𝗗𝗲-𝗗𝗲𝗺𝗼𝗰𝗿𝗮𝘁𝗶𝘇𝗮𝘁𝗶𝗼𝗻 𝗼𝗳 𝗟𝗶𝗯𝗿𝗮𝗿𝗶𝗲𝘀
𝗠 𝗥𝗼𝘀𝗲𝗻𝘇𝘄𝗲𝗶𝗴
The arrival of artificial intelligence in professional librarianship has created a sense of profound ambivalence inside the field.
On one side there is the promise of new tools that can speed up routine tasks and widen access to information; on the other side there is the gnawing suspicion that these same tools may be designed, intentionally or not, to replace the people who built the profession in the first place.
The anxiety is not merely Luddite unease. It grows from a century of experience in which every major technological shift-microfilm, online catalogs, full-text databases-was accompanied by claims that human mediation had become unnecessary.
What makes the current moment feel different is that AI is being marketed as a replacement for judgment, not just a supplement to workflow, and judgment has always been the librarian's deepest source of authority.
The fear of extinction comes from two converging tendencies. The first is managerial: administrators, trustees, corporate vendors, and university provosts have spent decades looking for ways to reduce labor costs in knowledge institutions. They hear the word "automation" and immediately imagine a budget without librarians.
AI vendors reinforce this by presenting their systems as frictionless substitutes for professional expertise. They boast that their products can classify, summarize, recommend, and retrieve information without human intervention. They present reference work as a series of disembodied functions rather than a social relationship in which a human being helps another human being confront uncertainty. When decision-makers accept this framing, the librarian becomes invisible, reduced to a cost center that can be trimmed once an AI subscription is purchased.
The second tendency is cultural: the ideology of technological inevitability that treats each new computational advance as destiny rather than design. Public discourse rarely acknowledges that technologies are shaped by politics, labor struggles, funding priorities, and corporate ownership. AI is treated as an unstoppable force that will absorb all cognitive labor, from legal analysis to clinical triage to the interpretive work at the core of librarianship. In such a narrative, resistance becomes a kind of naïve nostalgia, and the extinction of certain professions is described with the same tone as changing weather patterns. This fatalism spreads easily inside institutions where morale is already low, where years of austerity have left librarians wondering whether anyone still values their work.
Yet librarianship is not simply a bundle of tasks; it is a social practice that binds people to the idea that knowledge is a public good. Even the most polished AI cannot replicate this commitment. An AI system can find an article, but it cannot understand the stakes of a patron's situation. It cannot recognize that a question about eviction law may come from someone terrified of homelessness. It cannot sense when a teenager searching for information about gender identity needs more than just search results. It cannot read the emotional undercurrents in a reference interview, and it cannot serve as the trusted point of contact for a community that lives through surveillance, stigma, or exclusion. These forms of relational intelligence are not "soft skills"; they are the substance of librarianship as a democratic craft.
The deeper issue is that AI reshapes the landscape of information itself. Large language models rely on vast, scraped datasets whose provenance is often unclear, whose biases reflect the inequalities of the world that produced them, and whose errors can be presented with unwarranted confidence. To insert such systems into library service without human oversight is to turn the library into a distribution channel for unaccountable epistemologies. Librarianship has always insisted on traceability, context, and ethical responsibility. A model that generates fluent misinformation undermines these principles. The profession's survival will depend on insisting that AI be filtered through human discernment rather than the reverse.
This does not mean rejecting AI altogether. What threatens the profession is not the existence of the technology but the surrender of autonomy to it.
AI can help clean metadata, identify preservation needs, and analyze patterns in circulation or collections. It can help multilingual patrons navigate catalogs that were not built for them. It can accelerate technical services work in ways that free up human time for the interactions that matter most. But this requires librarians to define the terms of use. If libraries adopt AI as a tool shaped by human values rather than a replacement for human presence, the profession can expand its reach instead of seeing its core functions hollowed out.
The question of extinction is therefore inseparable from the question of power. Will librarians have the authority to shape the integration of AI, or will they be compelled to accept vendor-driven solutions that deskill their own labor? Will library administrators defend professional expertise, or will they treat librarianship as a workforce to be automated out of existence? Will the public understand the difference between cheap convenience and genuine access to knowledge, or will they be taught to think that a chatbot is the same thing as a librarian? The danger lies not in the code itself, but in the institutional willingness to erase the human relationships that libraries exist to nurture.
If librarians become extinct, it will not be because AI became too capable but because society allowed the logic of efficiency to override the logic of care.
Libraries would survive in name, but they would cease to be public spaces in the human sense. They would become transactional portals, optimized for metrics but stripped of the moral imagination that makes knowledge meaningful. The professional librarian-patient, political, attentive, resistant to the monetization of learning-embodies the counterweight to that future.
The challenge now is to articulate a vision in which librarians are not the victims of technological change but the authors of its ethical boundaries. AI will reshape the library whether librarians participate or not. The choice is between a future in which librarianship becomes a hollow brand attached to automated services, and a future in which librarians wield AI to strengthen their social mission. The second outcome is possible. Whether it happens depends on the profession's confidence in its own irreplaceability, not as clerks of information but as caretakers of the fragile, collective idea that knowledge belongs to everyone.