Where AI saves time and where it leads us off the map.

Over the past year, I’ve found myself replacing my favorite conversation starter—what are you reading?—with a new question: How are you using AI?

It’s not because I want to talk about AI more than I want to talk about books (I definitely don’t), but I’m curious to understand how people are actually using AI in real life. My husband reports that he’s started to use it at work to polish email messages, which he appreciates because he doesn’t particularly like writing tasks. I, on the other hand, imagine that prompting AI to write an email in my voice takes about as much effort as writing the email in the first place.

At the LJ Directors’ Summit last year, every vendor presentation touched on AI. Our podcast partners at Libraries Lead dedicate an entire segment—AI Watch—to the topic each month. Yet I remain somewhat befuddled about the ways I can ethically and responsibly use AI to make a positive difference in my work. Sure, AI is now embedded in standard office tools that many of us use regularly. I’ve found Otter AI to be helpful for a first pass on transcriptions, and Google Notebook LM is good for organizing notes. What I want to know is, how can I use AI to streamline the work that takes time but requires little critical thinking?

For instance, every month I spend a couple of hours scanning receipts, adding PO numbers, and uploading them to our expense system. I hate doing it, and it’s time I could otherwise spend reading articles, calling contacts, or brainstorming ideas for LJ coverage. Can anyone tell me how to AI away expense reporting?

Just a couple months ago I asked ChatGPT to help me with a task. I had a list of a few dozen cities that I wanted to see geographically distributed on a map. I uploaded the list, asked for the map, and ChatGPT enthusiastically responded that it was up to the job. When it offered a link to the requested map, there was no file to be found. “Are you sure you created a map?” I prodded the robot. “Ah, good catch! I wasn’t actually able to create that for you,” it replied. It then offered to produce a “simplified map,” which sounded just fine to me. What it generated was incomprehensible—a map of the United States but with cities incorrectly placed. San Antonio, according to this map, exists in Texas, New Mexico, and Louisiana; a city called “Los Aheles” is in California; and the mysterious “Rikehat” can be found in Virginia. Not helpful.

LJ’s 2026 editorial calendar has plenty of planned AI coverage—starting this month with Matt Enis’s feature story on AI policies in public libraries—and, I imagine, plenty of news stories about AI that will break. In our recent roundtable on copyright, AI dominated much of the conversation, which wasn’t entirely surprising, but it reinforced the need for LJ to stay on top of the legal and ethical issues related to AI.

Our team behind LJ & SLJ Professional Development has a slew of AI courses planned for 2026. Coming up this spring: AI Tools for the Overworked School Librarian (March 24), Teaching Ethical AI Research Skills (March 24), and an AI “lightning learning” session on March 4. You can find me at the February 26 session, AI for Librarians Who Don’t Have Time for This, where I hope to discover tools that are more reliable than the one that would send me to Rikehat.