I was fortunate to receive a WHSLA Professional Grant to attend the MLA '26 Conference held in Milwaukee last May. This post full fills the grant requirement to share what I learned with my WHSLA colleagues.
I was pleased to be able to attend MLA this year so close to home, and glad that WHSLA made it possible for so many of us to attend this year. [THANK YOU, WHSLA!]
Having attended several MLA Conferences over the last 25 years, this one seemed much reduced in terms of attendance and vendors. The first MLA conference I attended was in Chicago in 2000 (or so) and had roughly 5000 attendees in person. This time, it was less than 1000 people, with some attending virtually. Despite the reduced attendance, MLA still put on a great conference with many interesting presentation topics. Many attendees gave multiple presentations.
I did appreciate all the gouda cheese puns woven throughout the conference. Ha Ha!
This paper that stuck with me the most was:
Beyond the Widget Count: Telling Our Story with Microsoft Loop and Co-Pilot by Katherine Staab, AHIP, of Kaiser-Permanente (KP) Libraries.
Although I didn't know it when I sat in on this paper, it was a continuation of an older KP effort to show the value of libraries not just with statistics but also a narrative to explain the numbers. Their efforts have evolved to incorporate some AI tools, specifically Microsoft Loop and Co-Pilot, to collect the monthly narratives from ALL the KP Librarians to help fill in the bigger story about How much did we do in the last month? and What changed because of what we did? They also made an effort to link it back to company goals and objectives, which translates the value of Library Services into the language organizational leaders will understand. The tools make it much easier to gather the narratives from library staff so that 1 person can pull it into a collective report.
This immersion session stuck with me the most:
AI for the Uninterested: How I Learned to Stop Worrying and Use AI
The panel opened with an explanation of the AI Slop Cycle which starts with a hype cycle full of inflated expectations through the Trough of Disillusionment and finally to the Plateau of Productivity. It seems like we may finally be working through the cycle to find some actual productive uses for it now.
The second speaker made a case for Responsible Use of AI: It's ok to use AI tools in designer mode, to brainstorm ideas ...
Another speaker seemed to embody all the anxiety and doom wrapped up in AI. She was most concerned with the environmental and climate impacts of AI, using up water and energy. She brought up the dark side of AI that we never really hear about: toxic waste and heavy metals, as well as the human cost of mining for these heavy metals. She said a 30-second AI-generated video uses 1 hour's worth of energy. She advised attendees to stop making AI images and critically analyze what AI tools we do use. I had not thought about these issues before her presentation.
Another speaker said that medical students have real concerns about not only the environmental impacts, but also the risk of bias and discrimination in AI data. It's refreshing to hear that some medical students are reluctant to use AI tools, and are cautious about using it. This school directs their students to use subscribed and edited knowledge tools for trusted answers. They teach students to use the CRAAP Test to evaluate clinical vignettes with their students.
The message they try to convey to students is: Just because you can doesn't mean you should. Students seem to know that using AI is the wrong way, even though it's faster. Someone also made a point that ads in OpenEvidence pop up BEFORE the answer does, citing conflict of interest issues.
The last speaker focused on using AI for critical appraisal. She acknowledged that people struggle with reading research and interpreting statistics. She suggested using a generative AI agent (GemAgent in Gemini) to serve as a critical appraisal mentor to learn this skill. You can upload a pdf article and ask questions as you work through the evaluation. It's amazing, but I couldn't help thinking how this could go very wrong for a beginner.
This session was very well attended with people sitting at round tables to facilitate small group discussion. Some of the questions and responses were as follows:
1) Are there specific workflows where AI saved time in Libraries?
- AI Chat bots - contained and closed with a database made by librarians
- Communications and admin tasks such as scripts for similar questions. AI has been helpful in asking to soften the tone or clarify communications.
- Accessibility - where non-coders code
- Collection Development tasks -- Have it check Doody's
- CME Evaluations - ID themes
- Find things you're missing
2) What training can we offer about using AI Tools in Libraries?
- Prompt Engineering
- Comparisons between tools with breakdowns
- Evaluations
- When to use what when
- AI Literacy
- Gen AI, machine learning, data sets
There was a lively discussion at the end of the session. One medical librarian suggested that Librarians are the canaries in the coal mine: We can still tell what's real -- at least some of the time! And we're faster than some AI answers.
Librarians can train people to use some of these tools:
- Did it cite sources?
- Did you check the cites it referenced?
- Did the cites referenced say what the AI says it said?
- Use AI to learn, not for answers.
- Sometimes, it's easier and faster to use PubMed or the trusted databases because you don't have to double check everything.
Another big difference at this conference was a much reduced presence of The National Library of Medicine/National Network of Libraries of Medicine (NLM-NNLM) presence at MLA Conferences, and this time, the NLM Update was the last thing on the program Friday afternoon. Unfortunately, I missed it in an effort to beat the traffic getting home before a long holiday weekend. Fortunately, the session was recorded, and I was able to watch it later:
2026 NLM/MLA Joseph Leiter Lecture | How AI Is Reshaping Biomedical Discovery
In talking with the representative from UW-Milwaukee SLIS, I learned a bit about how the 2 Library schools in Wisconsin have evolved through the years. UW-Madison SLIS merged with the computer and engineering departments a few years ago. When I went to school there in the 1990s, the focus was public and school libraries and helping people find information--much more of a people focus. Now, the focus seems to be more on data and less on people. By contrast, UW-Milwaukee merged with the social sciences rather than engineering. This really surprised me because I always thought Milwaukee SLIS was more aligned with engineering than social sciences. Very different approaches to our subject matter. It will be interesting to see how these pan out ...
Those are the highlights from the MLA'26 Conference in Milwaukee. There was so much going on, I'm looking forward to what other WHSLA members learned.
No comments:
Post a Comment