Librarian — tasks, one by one
The unit of analysis is the task, not the job title. Each one below carries its direction, whether the judgement rests on evidence or on platform inference, the reasoning, and what it does not establish.
Every task on this page#
Helping people find what they need
Being augmented≈ Platform inferenceQuestions at the desk, by chat and in classes: working out what someone is really looking for, and where a trustworthy answer is.
Search engines and AI assistants now answer many of the quick factual questions that used to come to the desk. What remains is the harder part: working out what someone actually needs, knowing which sources to trust, and checking an answer a tool produced. The work is augmented rather than replaced because a wrong or invented source costs the person who asked, and a librarian is often the one asked to check.
This rests on what search and AI tools can do, not on a count of the questions libraries receive or how that number has changed.
Teaching people to use — and question — AI tools
New task≈ Platform inferenceClasses and one-to-one help on how search and AI tools work, how to check what they say, and how to spot an invented citation.
Libraries have long taught digital skills; generative AI adds a topic people now bring to the library themselves: how to use these tools and how to tell when they are wrong. It is new work created by the technology, and it draws directly on the reference librarian's habit of checking sources.
This page has no record measuring how many libraries run such classes or how many people come; it is an inference from the direction of the work.
Choosing and cataloguing the collection
Being augmented≈ Platform inferenceDeciding what the library buys or licenses, and describing each item so it can be found — subject headings, classification, metadata.
Much catalogue data already arrives ready-made from suppliers and shared records, and software can now suggest subject headings and summaries. Choosing what a particular community needs, and deciding when a machine-made description is wrong or misleading, stays with the librarian.
This rests on how cataloguing is done and what tools can suggest, not on a measurement of how much original cataloguing libraries still do.
Lending, returns and reservations
Automating≈ Platform inferenceChecking items out and in, handling holds, and keeping the self-service kiosks and return chutes working — work librarians share with, and supervise, library clerks.
Borrowing is the part of library work most thoroughly handed to machines: self-checkout kiosks, automated returns and sorting, reservation lockers. In 2023 Singapore's National Library Board opened a fully self-service 'Grab-n-Go' library in a mall, where patrons scan their ID to enter and borrow by walking out through a gantry, after piloting the same system at the National Library building the year before.
The mall library was a six-month pop-up with 5,000 manga titles; it shows the service can run without a desk, not how many libraries now do, and staff still stock, fix and supervise such a space.
Keeping the shelves and stock in order
Automating≈ Platform inferenceShelving, checking shelf order, taking inventory, and removing worn or outdated items.
RFID tags let staff take inventory with a handheld reader instead of item by item, and some libraries use robots that scan shelves for misplaced books. Deciding what to remove from a collection is still a judgement about what a community uses.
This rests on what the equipment can do; this page has no dated record of how many libraries use shelf-scanning robots.
Programmes, children and the community
Still human-led≈ Platform inferenceStory times, school visits, talks and clubs, help for people who come with a problem rather than a question, and keeping a busy public space calm and orderly.
This is the part of the job that is a room full of people: reading to children, running a class, noticing someone who needs a different kind of help. Software can advertise and book these programmes; running them, and judging who needs what, is done by staff.
This rests on the nature of the work, not on a measurement of attendance or of how staff time is divided.