Why your library’s AI future depends less on technology than on your boss!

Picture a librarian who spends her evenings teaching herself ChatGPT. She works out how it can help students track down sources, build reading lists, and find their way through tangled databases, and by the end of the week, she is ready to bring it to work. The next morning, she walks in, and nothing moves. The tools sit untouched, and the old workflows roll on exactly as before. That quiet stall is playing out across hundreds of academic libraries in India right now, and it points to something most technology stories miss: personal readiness counts for very little when the real gatekeepers sit somewhere else.

Researchers at SPJIMR went looking for those gatekeepers. They surveyed 497 academic librarians across India and analysed the responses using UTAUT, a well-tested framework for predicting whether people adopt a new technology. The model weighs four forces against each other, namely how useful a tool seems, how easy it is to handle, what the people around you expect, and whether your institution actually backs the change. The team then asked a sharper question than the model usually asks. They wanted to know whether those forces drive behaviour directly or whether they first must change how a librarian feels.

The answer surprised them, because feeling controlled almost everything. Of the four forces, the expectations coming from colleagues, supervisors, and the wider institution shaped attitudes more powerfully than usefulness or ease ever did. Librarians, in other words, read the signals around them before they decide anything, and the loudest signals come from the top rather than from peers. One respondent stated the problem without any hedging: “It should come from the higher authorities, and higher authorities are not willing to take it up.”

That single sentence overturns the usual script. We tend to assume that a tool which is easy and obviously useful will spread on its own, yet conviction at the desk keeps colliding with inertia in the corner office. The research clarifies the motivations and behaviours of librarians. Librarians act on AI only once they feel positively about it, and that feeling grows or withers according to the environment their leaders create. Leaders who stay neutral therefore choose, without meaning to, to keep their staff frozen.

This study found that people ready to adopt AI hold the least power to do so; front-line staff made up more than three-quarters of the survey sample. Front-line staff include the people who deal with users of libraries every day and spot first where a tool might help. Senior administrators, the ones who approve budgets and policies, accounted for under six percent and rarely see usage of tool on ground. Younger librarians were more willing to try new ideas and follow what others were doing. More experienced staff preferred to stick with familiar methods. As a result, enthusiasm for change often meets resistance, and caution usually prevails.

Geography complicates the picture further, since India does not behave as one uniform market. In the east, where budgets are thinnest, a positive attitude converted into genuine intent more strongly than anywhere else, almost as though scarcity sharpened the resolve to find something that works. In the north, by contrast, librarians weighed raw usefulness above how they felt. Any leader who imposes a single national plan will therefore watch it succeed in one region and stall in another.

The researchers refused to rest on their own numbers alone. They handed the findings to seven veteran librarians, each carrying between six and twenty-two years in the field and asked them to score the conclusions. The panel embraced the case for training and trustworthy output, awarding it well above eight out of ten, yet they split completely over money. One expert believed that financial support is a key barrier. Another thought it was a minor concern. Their different views, despite using the same evidence, show that context matters more than any universal solution.

Leaders need to step in and solve issues of limited funding and lack of training. Staff are more likely to accept change when they can see senior leaders actively supporting it. When leadership stays on the sidelines, frontline library staff may lose confidence and hesitate to move forward. Rather than just expressing support, institutions should provide the training, resources, and guidance staff need to make the initiative work. Respondents indicate subscriptions to ChatGPT as barriers to usage, thereby enhancing the productivity and quality of library services. Those who are able should engage directly. The interest that matches them already exists one level below them, and only their approval and, more visibly, their presence will unleash it.

Library employees’ GenAI adoption lesson has applications that go beyond libraries. We continue to think that technology is adopted and deployed the moment enough people warm to it, but in any cash-poor, hierarchy-heavy organisation, the staff are usually ready long before the institution and waiting for the right signal to proceed. The changing of a culture or environment, institutional change, is achieved by action, by just doing it.

The study titled “Attitude mediation in Indian academic librarians’ acceptance of generative AI: A UTAUT-based study of technology adoption in Indian academic libraries”, authored by Prof. Milind Kamat, M.F. Kumbar (Librarian at SPJIMR) and Uchita Bakshani (RA at SPJIMR), has been published in Elsevier’s Journal of Academic Librarianship.

Read more