Site LogoClose circle
News

News

Notes from Executive Director Scott Garrison – June 2024

Jun 3, 2024 | News

Continuing from my May MCLS newsletter piece on artificial intelligence (AI), this month I’ll recap Western Michigan University (WMU) Libraries Associate Dean Mary O’Kelly’s excellent presentation on using AI tools for research and writing at the recent Michigan Academic Library Association (MiALA) conference. Many facets of AI and how various industries are using AI tools show up in the news on a daily basis, and O’Kelly noted that current AI tools are affecting every aspect of education. In an interactive presentation that included poll questions she asked the audience via Mentimeter (which itself can group responses using AI), O’Kelly focused on three complementary AI tools that, used together, can search, analyze, and draft research articles: 

  1. Litmaps allows us to search Google Scholar for a topic and creates a mindmap connecting a set of related articles that shows number of references, number of times cited, and date for each. To reorient around an article’s references or citations, simply click on that article in the map and then Litmap’s references or citations function in the article brief summary view. Litmap’s color-coded tagging feature allows us to see articles we’ve tagged as we shift our view around a related set of articles via references or citations (where relevant). Litmaps also links out to article full text (subscription required in many cases), and can export a list of tagged references. Litmaps uses AI to analyze content for connections and pulls article metadata from Crossref, Semantic Scholar, and OpenAlex. 
  2. Elicit allows us to ask a research question and summarizes a set of articles in a top paragraph with citations. It displays the set in a table that can include user-selectable columns for basic citation information, summary, main findings, methodology, summary of discussion, and more. Elicit can also analyze PDF files the user uploads to it. Like Litmaps, Elicit searches academic papers from Semantic Scholar. 
  3. Once we’ve found relevant research articles, we can upload or paste full text and data into Claude, which can write an essay that compares and summarizes the information we provide it (and we can copy and paste the summary into other software for further editing). Claude can also create a bibliography in our preferred style. O’Kelly considers Claude the most powerful text analysis tool available. It was built by Anthropic (former OpenAI employees) to be more trustworthy and reliable than other tools such as ChatGPT.  

Having experimented a bit with ChatGPT and then Claude, I got curious about how each was trained. When I asked ChatGPT how it is trained, it replied: 

“I’ve been trained on a diverse range of internet text, including books, articles, websites, and other sources, to develop a broad understanding of human language. My training involves processing vast amounts of text data and learning the patterns and structures inherent in language, allowing me to generate responses to various prompts. If you have any specific questions about my training or how I work, feel free to ask!” 

When I asked Claude how it is trained, it replied: 

“I was trained using constitutional AI, which involves using machine learning techniques to align my behaviors with certain guidelines or principles during the training process. From what I understand, the key aspects of my training approach were: 

  1. Oversight and Curation: The training data I ingested was carefully curated and filtered by the Anthropic team to remove potential sources of bias, toxicity, or misinformation. 
  2. Constitutional Rules: A set of key principles around being helpful, harmless, honest, ethical, and protecting individual privacy were encoded into the reinforcement learning that shaped my behavior. 
  3. Multi-Task Learning: In addition to open-ended dialogue, I was trained on a wide variety of tasks like analysis, coding, math, and question-answering using a diverse corpus of high-quality data sources. 
  4. Iterative Refinement: My outputs were continuously evaluated and my training was refined through techniques like reinforcement learning and recursive reward modeling to better align me with Anthropic’s intended principles. 
  5. Knowledge Cutoff: My training data cutoff in August 2023, giving me a stable and consistent knowledge base that doesn’t shift over time. 

However, many of the technical details like model architecture, computer resources used, hyperparameters, and the precise data sources involved are not known to me. I aim to be as transparent as I can about my training while respecting the intellectual property of Anthropic. Please let me know if you have any other questions!”  

Noting Claude’s transparency and being curious for more about ChatGPT, I also asked Claude how ChatGPT was trained. It replied:  

“I don’t have insider knowledge about the specific training process used for ChatGPT, as that is proprietary information held by OpenAI. However, based on publicly available information, we know some key points about how ChatGPT was trained: 

  1. Constitutional AI: ChatGPT was trained using an approach called “constitutional AI” developed by researchers at OpenAI. This involves using reinforcement learning to instill desired behavior in alignment with set of rules or guidelines. 
  2. Web Data: A substantial portion of ChatGPT’s training data comes from crawling and ingesting a massive corpus of text data from the internet, including websites, books, articles, and other online content. 
  3. Supervised Fine-Tuning: After pretraining on the broad web data, ChatGPT went through additional supervised fine-tuning on smaller, curated datasets to enhance its abilities in areas like question-answering and dialogue. 
  4. Reinforcement Learning: Techniques like reinforcement learning from human feedback were used to refine ChatGPT’s outputs to be more relevant, truthful and harmless. 
  5. Model Scale: ChatGPT leverages an extremely large neural network model with over 175 billion parameters, allowing it to ingest and learn patterns from a staggering amount of training data. 

However, many details about the model architecture, computer resources used, hyperparameters, techniques to instill certain capabilities, and the exact datasets involved remain closely guarded by OpenAI. Their published research only discloses high-level methodologies. But this outlines the general training approach at a high level from publicly available information. Let me know if you need any other perspective on this topic!” 

When I asked Claude to compare its own training with ChatGPT’s, it told me the following based on publicly available information, which I’ll summarize here: 

  1. ChatGPT was primarily trained through crawling a large number of Web pages, and Claude’s training was curated and filtered by Anthropic. 
  2. Both were trained using constitutional AI, with Claude’s training instilling principles around being helpful, harmless, honest, and ethical. 
  3. While ChatGPT’s training data is updated periodically, Claude has a training cutoff of August 2023, which it says provides stable and consistent knowledge. 
  4. Claude has undergone multi-task training across areas like analysis, coding, math, and question-answering, and to Claude, ChatGPT’s training “seems more focused on general language skills and open-domain knowledge.” 
  5. Finally, whereas ChatGPT uses a 175 billion-parameter model, Claude’s model “is comparatively smaller in scale.” 

Although Claude can hallucinate like other LLMs do, users can upload the exact text they want to analyze, minimizing errors. Claude also has been programmed to say, “I don’t know,” and to respond only if confident in its response, if prompted to do so. Claude can analyze pasted text and uploaded Excel .cvs files, Word documents, PDFs, and .txt, .rtf, and .epub files. It can also be embedded into Google Sheets to provide even more data analysis help. ChatGPT told me that I could paste text into a conversation with it, or provide it links to publicly available content.   

O’Kelly briefly mentioned some other AI-based tools such as the Packback writing/grading assistant, Clarivate’s Alethea deep reading tutor (see also a May 21 press release for more about Clarivate’s AI strategy), and ResearchRabbit, and noted that Zotero now has ARIA, an AI-based plugin. She also touched on what some tools purport to do versus actual mixed results, and some tools’ false alarms when trying to detect AI in student work. In closing, O’Kelly emphasized that vendors are putting AI into their systems that libraries and their users will need to understand, and that faculty and students are using tools like she described. She asked the audience to consider what all of this means in terms of library discovery, how users use library collections, and the library’s role. O’Kelly suggested that session participants watch for more on the paradoxes of AI in academic libraries from a chapter she is writing for the forthcoming book AI for Academic Libraries: Navigating the Future. 

My thanks to Mary O’Kelly for her enlightening presentation, and her guidance on this summary.  

What are you doing with AI currently, and what do you think about it? Let me know, at garrisons@mcls.org. 

Related News

Featured Image

MCLS welcomed two new Administrative Assistants in September 

On Sept. 28, 2026, MCLS welcomed Renea Di Bella and Manaal Kahn as our new Administrative Assistants. You may engage with them through RIDES and MelCathelp tickets, when you call our main phone line, and when they host our virtual workshops and webinars. Among other duties, they also help process membership forms and contact management data. When you have the pleasure of interacting with them, please extend a warm welcome to MCLS! In the meantime, the interviews below provide deeper insight into Renea’s and Manaal’s backgrounds and aspirations.

Meet Renea Di Bella

Tell us a little about your background. How did you get started in libraries?

I am a former middle school social studies teacher with a master’s degree in curriculum and instruction. Through writing curricula, I realized a passion for literacy, research, and archival preservation. To pursue this passion further, I applied (and was admitted) to the MLIS program at Wayne State University, to begin Fall 2027. In the interim, I started working in libraries to prepare myself for my studies and future career. I worked previously at the Detroit Public Library and am excited to bring my passion and interests to MCLS.  

What motivated you to take on this new role?

This role merges several of my interests and talents. Teaching provided me with many of the skills I will need to fulfill my role, while the services MCLS provides its members align with my passion for literacy and public access to information. This is also an opportunity for me to get to know the needs of libraries better as I transition into a career in Libraries and Information Science.  

What do you like to do when you aren’t working?

I like to read and travel. I like to do several different types of art and crafts. I also spend an inordinate amount of time just watching my cats exist. 

Do you have any hidden talents or hobbies?

I am a visual artist; I make multimedia collages and sculptures. 

What is the best advice you have ever received?

From my aunt: “Don’t dread getting older because it’s a privilege not everyone gets.” 


Meet Manaal Khan

Tell us a little about your background. How did you get started in libraries? 

I worked in the Processing department at the Dearborn Public Library, where I supported MeLCat requests by printing pull lists from the ILS, retrieving materials from shelves, printing RIDES labels, and preparing items for RIDES delivery. I also processed materials for circulation, and my favorite part was processing holiday materials! My background also includes instructional design, literacy tutoring, and technical writing, where I focus on how to break complicated information into clear steps and help people apply what they learn. 

What motivated you to take on this new role? 

I am motivated to bring together my library experience, communication skills, and interest in helping others apply what they learn. Working with MeLCat and RIDES showed me how libraires work together to connect people with resources, and I am happy to support that work at MCLS. 

What are you most excited to start working on? 

I am excited to combine my previous experience with RIDES and MeLCat. I value literacy and libraries make literacy accessible to all. MelCat is an amazing part of the library network, giving people access to more resources than just their local libraries. I’m excited to support member libraries and make information accessible. As I grow, I’d love to use my instructional design and writing skills to help make instructions, resources and training easier to follow and apply. 

What do you like to do when you aren’t working?  

I enjoy learning languages and different subjects, and geographic information systems (GIS) is my next area to explore. I like to spend time with my nephew, play Zelda, sketch places in my neighborhood, read, cook, and watch anime and fantasy shows/movies, like the current Percy Jackson series. I also enjoy going on walks and taking photos of trees. 

Do you have a favorite author, book, or genre?

My current favorite books are Days at the Morisaki Bookshop by Satoshi Yagisawa and The Invisible Library series by Genevieve Cogman. 

Nov 2, 2026

Read more
Logo

Congratulations to our newest Certificate of Library Staff Development recipients (November 2026)

The following library staff members have earned Certificates of Library Staff Development by attending workshops through MCLS!  

Basic Essential Workplace Skills Certificate  

  • Tresa D. Fore, Wayne County Community College District, MI 
  • Terri Hebert, Bridgman Public Library, MI 
  • Bailey Jo Good, Wabash Carnegie Public Library, IN 
  • Eliza Forrest, Traverse Area District Library, MI 
  • Alisia Neubeck, Grand Valley State University, MI 
  • Christina Meyers, Traverse Area District Library, MI 
  • Lynnae M Christensen, Traverse Area District Library, MI 

Are you interested in obtaining your Certificate of Library Staff Development? For details, see our website. Essential Workplace Skills tutorials are free to MCLS members as a member benefit! For questions, please email Gwen at training@mcls.org or call 517-220-9950.  

Nov 2, 2026

Read more
Featured Image

Insights from Executive Director Scott Garrison – November 2026

Libraries are built from stories, and they embody the power of stories. Stories are at the heart of the relationships libraries and everyone they serve need to succeed. At the MCLS annual member event on October 9, approximately 70 total attendees focused on how libraries can tell stories effectively from the perspectives of data storytelling and crisis communication.

The path to MCLS’s first in-person event in ten years is made up of many stories from the astonishing number of changes and challenges libraries have seen since 2016. We at MCLS have collected a lot of data in the last few years that are helping us tell new stories, from our strategic plan development and implementation through recent surveys and other work our Board and staff have done.

Following some stories from MCLS Board members Mike Hawthorne and Josh Petrusa that set context for our event, Professor Kate McDowell from the University of Illinois Urbana-Champaign’s School of Information Sciences delivered an inspiring keynote based on her ten years of work on data storytelling including the Data Storytelling Toolkit for Libraries (DSTL) and her 2025 book Critical Data Storytelling for Libraries: Crafting Ethical Narratives for Advocacy and Impact. Kate’s book weaves together many helpful strands that can help us build a positive future, and I recommend it very highly to anyone seeking to improve how they combine data with narrative to tell their library’s stories.

In her keynote, Kate highlighted the Storytelling Triangle between a story, its teller, and its audience, noting that we should be mindful whose role it is to tell a particular story. Each relationship in the Triangle is a dynamic one that can be influenced by whether a story is new or being repeated by the same or a different teller, and an audience’s attitudes toward the story. Kate established that critical data storytelling “requires understanding power dynamics from a storytelling perspective,” an idea dating back to the late-19th century women who introduced using empirical data in librarianship in a systematic way, and who were not allowed to present at conferences at the time.

Kate pointed out the differences between data, story, and information with the straightforward idea that “data in story context becomes information.” In her keynote (and in much greater detail in the book), she offered thought-provoking ideas about who owns data (and goals for using data) versus stories (and motivations for telling them), with examples relevant to academic and public libraries. Kate laid out the Data Storytelling Life Cycle as a framework to dealing with struggles libraries have had including translating data into stories and capturing those stories for outside audiences.

Data Storytelling Life Cycle (McDowell, 2026; used by permission) [size this image so it’s big enough to read online and scale it correctly]

The Life Cycle captures the ideas that many libraries collect and store data without really engaging and creating meaning, and that we need to make our data our own through all four steps in the cycle.

Kate then offered suggestions for to how to build stories that last and should be retold over time, such as how libraries have ended late fees to promote equitable access. She explained the cycle of story across time, from telling, to listening, to retelling, as well as a story’s chronological structure from an inciting moment though rising action, to a climax or turning point and falling action to a resolution or denouement. She cited the example of a campaign to expand a taxing district for a library whose district was smaller than its counterpart school district. Though the campaign failed, the data the library used to tell their story along the “upward arc” of rising action in their campaign increased awareness that kids outside the taxing district could come to and use the library even without a library card.

Before closing, Kate laid out three Structure Narratives from the DSTL and pointed out where data collection occurs in each:

  • Transformation – a “hero’s journey” toward overcoming obstacles and creating change
  • Continuity – restoring stability following disruption (with recovery, adaptation, and learning)
  • Discovery – understanding complexity via a process of asking questions, gathering information, and building and sharing the knowledge that results from the process

Next, communication expert Thom Fladung led a workshop titled Putting Crisis Communications In Play, to help libraries be prepared in case of “explosive, unplanned visibility” that can damage our most important asset: our reputation. Thom’s experience as an investigative reporter earlier in his career has positioned him very well in his current role as Vice President and Chief Counselor at Hennes Communications (a division of AKCG Public Relations Counselors).

I hope you find the following summary of what Thom shared on October 9 useful. Rather than summarizing all of the valuable information he offered, please see below for images from a pocket-sized card with tips and more that Thom distributed and very kindly granted us permission to include in this piece.

Thom began with the idea that while communication can be the last thing we think about during a crisis, it can save our reputation as word spreads and the community awaits our response. He emphased how important it is to act promptly and communicate clearly and honestly in a crisis using the rules of the Damage Control Playbook:

  1. Tell the truth (“no comment” can translate to “guilty”)
  2. Tell it first
  3. Tell it all (i.e., all that you know to be true)
  4. Tell it fast
  5. Tell it to the people who matter most

Thom shared stories from his journalism experience to illustrate how the media works faster and more iteratively than ever before in the age of social media and search engine optimization. Media often run with what they have even if it does not represent a balanced view of a story.

He then turned to a set of five steps to manage outrage that often occurs in a crisis:

  1. Validate concerns: there is almost always a valid concern to acknowledge
  • Acknowledge past mistakes: don’t pivot away from the past more quickly than your stakeholders can – it’s their decision when to move on, not yours. Be prepared to repeat how you’re working to fix past mistakes, if needed.
  • Acknowlege current problems: iflibrary leadership does not demonstrate that they know there’s a problem, stakeholders may become upset
  • Give credit: wherever possible, thank anyone who is helping resolve a crisis
  • Share control: the higher the outrage, the less trust there will be in you to solve a problem – let others validate whether you’ve solved it

Thom laid out a “3-V framework” for bad news stories: the villain who caused the problem, the victim, and the vindicator who brings it to light for accountability. He also offered detailed strategies and tips for communicating via social media, pointing out the importance of having conduct rules for all communication including social media and how to handle comments effectively. Thom cited Traverse Area District Library’s Social Media Policy as a great example. In a crisis, he recommended putting the truth on the library’s website and making a simple social media post pointing to the website because a library has greater control over its website. To help defuse a social media storm, ask vocal commenters for a conversation away from the platform.

Libraries should ask the following questions when they identify an online reputation threat:

  1. Where did the threat originate?
  2. Are more people continuing to talk about this situation?
  3. Has the mainstream media begun covering this situation?
  4. What is your understanding of what led to negative information?

Following tips on determining the influence of a threat and best practices for responding, Thom turned to some media relations guidelines and interview control tools if we find ourselves contacted by the media. He staged some mock media interviews with brave event participants, seeking observations from the larger group.

Thom closed with key principles for developing a crisis communications plan:

  1. Ask “who says what to whom, how and when?” Who is on your library’s crisis communication team, and what protocol/procedures will you follow to activate that team?
  2. Conduct a vulnerability audit with the library leadership team, brainstorming and ranking potential crises by how likely they are
  3. Build messages from the audit, including media statements, website content, and social media posts
  4. Make sure you identify your key stakeholders and audiences, know your social media instructions, passwords, and guidelines, and keep contact information up to date

As promised, here is the card Thom offered workshop participants:

© Hennes Communications LLC; used by permission

I’ll close here by thanking our annual member event planning committee, the staff and Board members who played a part in making the event happen, our hosts at St. Joe County Public Library, our fine speakers, and all member library staff who attended (especially those who completed our evaluation survey). Holding our first in-person event in many years was a lot of hard work, and truly a labor of love for our members.

If you were with us on October 9, I hope you’ll agree that our event was a great example of how MCLS brings libraries together to engage with valuable strategies and approaches, and of how together, we empower a strong library community. I also hope to see you at our next annual member event on April 30, 2027, back at St. Joe County Public Library in South Bend, IN. Until then, I encourage you to share questions, ideas and thoughts with me at garrisons@mcls.org.

Nov 2, 2026

Read more