Connecting AI and ML to Libraries
I am very appreciative for the opportunity to attend the IDEA Institute. It has been wonderful listening to the ideas about applying Artificial Intelligence (AI) and Machine Learning (ML) to librarianship. Often, people do not associate librarianship with AI and ML. However, the disconnect is a misconception. Libraries are already using AI and ML and the applications of it will only continue to grow as librarians embrace AI and ML and adapt services.
Considering Coding and How It May Impact Project Results
We had a presentation from Jason Griffey (2019) and watched the movie Coded Bias. Mr. Griffey shared several ways that AI and ML are currently being used and indicated some of the dangers that are associated with them. Similarly, Coded Bias described how algorithms included unfair biases. Yet, there is no regulations of the information that they collect or how they are designed. Algorithms are already impacting the social services we receive, the way we shop, college applications, credit, and much more.
Regardless of the project that we decide to do, we much always be aware of how our errors, biases, and planning can influence the quality and outcomes of projects. Moreover, each type of project can present its own contextual obstacles and shortcomings. We must strive to make ethical decisions and to promote access, equity, inclusion, and diversity.
Project Planning
In our planning, we should consider how we consult with others to create the best products. AI and ML projects start with research and synthesis. These processes look chaotic and as we move to the concept and protype stages, the chaos begins to dissipate. During the institute, we were taught to use the following steps for project design.
1. Identify the problem: School librarians need access to specialized professional development.
2. Inventory resources: I have an assistant to help me. In addition, I have a community of practice for feedback (includes faculty in various subject areas, graduate-level information science and data science students, subject level experts, current and pre-service school librarians), and access to research. I will ask for help from other faculty and my new IDEA network if needed.
3. Review methods: I learned that Scikit-learn and SPACY can help me with using Natural Language Processing to analyze articles and the presentations collected at a symposium. The articles and presentations have strategies that school librarians can use. There are web scrapers that can be programmed without codes.
4. Define Success: To begin, I will have a chatbot that answers at least 300 common questions from school librarians. The project will continue to develop, and I will retrieve new questions from the chatbot queries. The chatbot and tutorial will be reviewed by students, school librarians, and professors to determine their usability. After they are approved, they will be placed online.
My project evolved. However, learning Natural Language Processing is on my list of priorities. I believe it will help me with my research. It was suggested that I read Samuel Burns’ (2019) Natural Language Processing: A Quick Introduction to NLP with Python and NLTK. I also learned that I may not need to train a model for my data analysis. However, ML has risks and I should review the results with other experts. As such, the presentations that were delivered at the symposium probably will need specialized attention, depending on the results. Hugging Face has pre-trained models that can be used for analysis.
My Project
I had many ideas for projects. Ultimately, I decided to make a chatbot to serve as an on-demand professional development resource for school librarians. I plan to use common questions that current and pre-service school librarians ask to train the chatbot. Although there are several listservs that school librarians can consult and professional development offered by many organizations, these resources are scattered across the Internet. The chatbot can answer questions and direct users to resources in a centralized location.
Eventually, I would like to expand the project. I can create a tutorial for the chatbot and place it on a website with examples of other MI and AI applications for school librarians. I hope school librarians will look at these examples and consider implementing AI and ML activities for programming.
During the program, I learned about SMARTIE (Strategic, Measurable, Ambitious, Realistic, Time-Bound, Inclusive, Equitable Goals). Here is the SMARTIE Goal developed for the project. Create a free chatbot based on real-world examples of professional questions from school librarians. Develop a tutorial discussing directions for creating similar chatbots and identifying the application of ML and AI in school libraries by January 2022. Throughout the process, I plan to seek the feedback of various faculty, information science students, and the school library community from diverse locations, work backgrounds, and demographic representations.
I am interested in this project because the school librarianship community needs to examine AI and ML more. We teach our students to be safe online. AI and ML are addressed in coding clubs, makerspaces, data literacy, data leadership, and digital citizenship. However, our students need to understand the implications of AI and ML in their lives. Yet, AI and ML are often ubiquitous; although there are privacy, ethical, and equity issues.
I want school librarians to know that they can use ML and AI. I am thinking about how to teach about AI and ML in a user-friendly way. I will look for more applications such as Teachable Machine. I plan to explore MonkeyLearn as well. Tools like MonkeyLearn may have an economical solutions for educators to explore.
School librarians can also utilize AI and ML to introduce students to STEM careers. Democratizing STEM education is an immense opportunity for school librarians. I look forward to delving deeper into the implications of AI and ML in school librarianship.
References
Burns, S. (2019). Natural language processing: A quick introduction to NLP with Python and NLTK. United States: Amazon KDP Publishing and Printing.
Griffey, J. (2019). Artificial intelligence and machine learning in libraries. Chicago, IL: American Library Association.
