One fine morning, a young lady called me. Her voice carried anxiety. I assumed it was the usual exam pressure. I told her honestly that I was not a psychologist and might not help her deal with anxiety. She stopped me immediately.
“No sir,” she said. “It is not about the exam.”
She told me she was in her final year of Library and Information Science. She had been watching my videos where I explain how artificial intelligence makes a researcher’s life easier. Faster discovery. Instant summaries. Easy access to information. Then she asked a question that stayed with me long after the call ended.
“If researchers get information so easily,” she asked, “who will come to libraries? And if nobody comes, who will hire librarians like me?”
That single question captures the fear many students, early-career professionals, and even senior librarians are quietly carrying today. It deserves a clear, practical answer.
To understand this properly, we need to slow down and separate hype from reality.
Artificial intelligence is powerful. It can summarise books, answer questions, and even draft research papers. It speaks confidently. That confidence often misleads us into believing the answers are correct. This brings us to the first and most important lesson.
AI gives answers.
It does not give judgement.
An AI system does not know whether information is biased, incomplete, outdated, or ethically problematic. It predicts text based on patterns in data. If the output sounds fluent, the system considers its job done. Truth, context, and consequence are not part of its thinking.
Judgement still belongs to humans. And judgement has always been at the core of librarianship.
This leads to the second lesson, one many people underestimate.
Traditional library skills are not outdated.
They are AI-era skills.
Take cataloguing. Many students see it as mechanical and irrelevant. In reality, cataloguing is structured thinking. It is about describing information so others can find it, understand it, and trust it. Today, AI systems depend on exactly this kind of structure.
AI models need clear documentation.
They need clean metadata.
They need transparency about data sources and limitations.
Without these, AI becomes a black box. Librarians have been preventing black boxes for decades.
The same applies to information retrieval. Long before AI existed, librarians taught users how to search effectively, refine queries, evaluate sources, and understand context. Modern AI search works well only when someone understands relevance and authority. That skill has not disappeared. It has become more valuable.
Then there is ethics.
Libraries have always stood for access, equity, privacy, and intellectual freedom. These are not optional values in the AI age. They are essential safeguards. AI systems amplify bias, exclude voices, and compromise privacy if left unchecked. Librarians already know how to question systems, not worship them.
This is why an important shift is taking place.
Librarians are no longer only users of AI.
They are becoming the human infrastructure behind AI.
They ensure systems are transparent.
They ensure systems are fair.
They ensure systems serve people, not mislead them.
This is not a future scenario. It is already happening.
A 2025 Clarivate report shows that 67 percent of libraries are already exploring or actively using AI. Libraries now operate in a research ecosystem where AI tools scan thousands of papers, extract data, generate cited answers, and map research connections visually.
These tools save time. They also confuse users. Researchers often do not know where answers come from, what was excluded, or what assumptions were made. Someone must explain this clearly.
That responsibility naturally falls on librarians.
Behind the scenes, AI is also reshaping library operations. Metadata creation, cataloguing, and collection management are increasingly automated. A system can generate records. A model can catalogue a book from an image. This does not remove librarians from the system. It removes repetitive labour.
What replaces it is higher-value work.
Advanced research support.
Teaching AI and information literacy.
Community programmes.
Policy guidance and ethical review.
Another fear needs addressing.
Many people assume AI will reduce the importance of libraries. In practice, it often expands access.
In India, mobile AI labs travel to remote villages. They do not replace libraries. They work alongside traditional village libraries. Technology moves, but trust remains local. Libraries become bridges between advanced tools and real communities.
At the same time, we must speak honestly about AI’s weaknesses.
One term everyone must understand is AI hallucination. This occurs when a system produces fluent but false information. There is no intent to deceive. Accuracy is sacrificed for smooth language.
The consequences are serious. Researchers have wasted hours chasing references that never existed, created entirely by AI. Proving that a source does not exist takes time and energy away from meaningful work. This feeds what many experts now call the slop problem, where low-quality AI content floods the internet and academic publishing. Trust erodes. Reviewers burn out. Good research gets buried.
So the practical question becomes unavoidable.
Why does AI still need librarians?
Because someone must teach critical evaluation.
Because someone must audit bias.
Because someone must protect privacy.
Because someone must identify fake citations.
Because someone must uphold intellectual freedom.
AI does not understand these responsibilities. Librarians do.
This brings us to the transformation of the profession.
The librarian is no longer a gatekeeper of information.
The librarian is a supervisor of AI systems.
The librarian is no longer only a reference desk expert.
The librarian is an AI literacy educator.
The librarian is no longer only a collection manager.
The librarian is an ethical evaluator of everyday tools.
The most accurate description of this role is information architect. Someone who designs, audits, and oversees how knowledge is created, accessed, and trusted.
This transformation requires investment. Not only in technology, but in people. The AI-ready workforce will be built, not bought. It will emerge through reskilling, confidence building, and empowering professionals who already understand information deeply.
When I think back to that anxious student, I no longer see a profession in danger. I see a profession at a turning point.
AI delivers answers faster than ever.
But society still needs someone to teach how to question those answers.
That responsibility has always belonged to librarians.
And it still does.