AI Library Mentor | Tereza Kiselová

The project focuses on using AI to personalize library resource recommendations for university students so that they can more effectively expand their knowledge based on their interests. The goal is to create an AI library mentor that analyzes user data—search history, borrowed items, and information about how the student wants to develop—and then provides targeted recommendations for library resources.

The solution takes the form of a chatbot with analytical capabilities that remembers the student's decisions and provides appropriate recommendations when prompted by the student. It includes a rating system so that users can indicate whether the recommended resource actually suits them. The project includes the option to exclude sensitive loans from the analysis to protect privacy and ensure accessibility for all students, regardless of their technological skills.

The result is a functional prototype created using Lovable, a test chatbot implemented in Gemini Gems connected to the MU library catalog, a dashboard with data analysis on student library needs, and a comprehensive ethical analysis using AI Ethics Canvas. The project confirmed that a similar specialized library chatbot does not yet exist in libraries and has real potential.

Ethics: The project comprehensively addressed ethical aspects using Tarot Cards of Tech and AI Ethics Canvas. It identified key risks such as privacy protection for sensitive topics (The Backstabber), ensuring accessibility for all students (The Forgotten), and preventing algorithmic discrimination (The Scandal). Proposed measures include the option to exclude items from analysis, transparent communication about how AI works, multilingual support, regular bias checks, and the option to contact a librarian.

AI TOOLS USED IN THE PROJECT

1. Gemini 2.0 Flash (free) – brainstorming, idea generation, basic data analysis, creation of Gemini Gems chatbot
2. Perplexity + Deep Search (free) – topic research, questionnaire and answer generation, project reframing, competition verification
3. Consensus (free) – academic research, searching for relevant sources on a topic
4. NotebookLM (free) – working with academic sources, summarizing documents, answers based solely on uploaded sources
5. Napkin.ai – creating visualizations and infographics for presentations
6. Google AI Studio – testing voice chatbot, streaming real-time features
7. Descript (free version) – transcribing audio recordings from interviews
8. PowerBI – manual creation of a dashboard for data analysis
9. Lovable (free) – prototyping a functional AI library application
10. Gemini Gems – creating and testing a personalized library chatbot
11. Canva AI – automatic creation of presentation designs

OUTCOMES AND PROTOTYPES

• Prototype in Lovable
• Gemini Gems test chatbot (integrated with the MU library catalog)
• Data analysis dashboard (PowerBI)
• Visualization from Napkin.ai
• Project presentation
• Research outputs from Perplexity and Consensus
• AI Ethics Canvas analysis (complete ethical framework in Miro)
• Audio recordings from testing (processed via Descript)