Notes on Information Literacy, AI Literacy, and Library Practices for the Course
Course Context and Presenter
Presenter: Jennifer, affiliated with the University of Michigan, focusing on information and library information services, master’s in clinical mental health counseling, and digital literacy for undergrad and grad students.
Roles and responsibilities:
Part of the central team supporting English and communication courses.
Manages library programming for first-year communication.
Teaches and conducts virtual classes; develops tailored courses and modules in Canvas.
Tracks and coordinates international outreach, building partnerships between the libraries and other campus units and beyond.
Class setup:
Jennifer was invited to lecture early in the semester to introduce the course before semester-long material.
Assignment structure announced: three assignments throughout the semester.
Assignment 1: Pick a topic, find a journal article from a reliable source; task to determine what source is reliable.
Assignment 2: Summarize the article, concept, or theory.
Emphasis on digital content creation and evolving topics (AI, generative AI) relevant to the course.
Contact and interaction:
Jennifer’s DT email address provided on Blackboard; students encouraged to reach out with questions.
Instructor also available after the presentation; class encouraged to ask questions during and after the talk.
AI, GenAI, and Library Context
Topic framing:
The session centers on GenAI (generative AI) and digital content creation, with emphasis on reliability, source evaluation, and ethical use in academic work.
How AI is portrayed in the talk:
AI uses crawlers/bots/scrapers to gather data within parameters from various domains (math sites, languages, medicine, etc.).
Content behind paywalls may not be accessible by AI, creating gaps in AI-generated outputs.
Collected data is blended into a large pool and used to generate answers that are summarized for user queries.
Strengths and limitations of AI:
Pros: fast, convenient, can automate tasks, reduce human error in repetitive data editing, scalable support (24/7) for many users.
Cons: not always complete (paywall content may be missing), not truly original (often remixing existing content), can mix or collide ideas in ways that are not coherent or accurate, may hallucinate or misattribute sources, privacy concerns due to user profiling.
Privacy and security concerns:
AI can track user behavior and create profiles, raising privacy worries.
Data sharing practices of large platforms (e.g., Meta) regarding data use and third-party sharing were discussed as context for reliability and privacy.
Practical implications in libraries:
AI can assist but may propagate incorrect or misleading information, which is problematic in high-stakes contexts (e.g., medical information for patients).
Librarians emphasize critical evaluation when using AI outputs for scholarly work.
Limitations and debate:
Some researchers argue GenAI cannot generate truly original material; it primarily merges/remixes existing content.
Genre-specific debates: e.g., attribution of novelty in AI-produced content; the idea that AI outputs are not original in the traditional sense.
Notable examples given:
Misattribution or misidentification of journals (e.g., confusion around titles like the Journal of Classical Sociology) to illustrate reliability concerns.
Real-world demonstration of variability: different AI outputs for the same prompt (the gummy bears in the Atlantic Ocean and related numeric prompts) to show inconsistencies across models.
Ethical and professional cautions:
When using AI in coursework or professional contexts, it’s important to verify outputs with trustworthy sources, cite appropriately, and avoid presenting AI-produced material as original authorial work.
In health or safety-critical contexts, avoid relying on AI for definitive information without validation from authoritative sources.
AI Output Demonstrations and Critical Thinking
In-class AI outputs examples:
A set of prompts about a hypothetical question (e.g., how many gummy bears fit in the Atlantic Ocean) yielded wildly different numerical answers across AI instances:
Example outputs included numbers like and (representing different scales or units depending on the model).
This illustrates how the same prompt can produce inconsistent results due to model internal parameters, prompting users to critically evaluate and cross-check results.
Environment/trade-offs visualization:
An app demonstrated energy use comparisons between activities (e.g., AI tasks vs charging a smartphone) with units described as energy consumption in watts per hour, illustrating the potential environmental cost of AI workloads.
Note: these were described as guesstimates, not precise measurements, and highlighted that more complex prompts requiring multi-step reasoning can shift energy use considerably.
Real-life risk considerations:
In medical or health contexts, inaccurate AI outputs can have serious consequences; practitioners must verify information with reliable sources before acting on AI-derived summaries.
Audience engagement:
Emphasis on interactive class discussions and critical questions about reliability, source quality, and the ethical implications of AI-assisted research.
Information Literacy and Source Evaluation
Discovery and search tools in the library:
Discovery search: Library homepage feature; comprehensive search across books, DVDs, and electronic resources.
Databases: Focused electronic resources for specific topics (e.g., sociology databases, psychology databases); typically more specialized than a general search.
Google and Google Scholar cautions: Similarities to AI in background data gathering; paywalls can limit access to full texts; library access often provides paid content at no direct cost to students.
Privacy-aware search tips: Use incognito mode or privacy-focused search engines like DuckDuckGo to minimize profiling; Google Scholar may not access paywalled content provided by the library.
High-cost access caveat: Some articles or journals may require pay-per-article fees if not covered by library subscriptions (e.g., engineering databases may cost hundreds of thousands per year to maintain).
Library mission and cost efficiency:
The library spends millions annually to provide access so students do not pay per article; emphasize using library materials to be cost-effective.
Information evaluation framework (SIFT): universal method beyond traditional journal credibility
SIFT steps (conceptual): Stop, Investigate the source, Find better coverage, Trace claims to the original source.
Application: Not limited to recognizing journals; useful for evaluating any claim, source, or article.
Clark Library guide referenced as a comprehensive resource with tabs and videos detailing SIFT and other evaluation strategies.
Additional evaluation tools and concepts:
Check journal and source transparency: About Us pages, ethics statements, and adherence to professional standards.
Retractions and updates: Look for corrections or retractions indicating the source’s transparency.
Distinguish between news and opinion, and be cautious with emotionally charged content (satire, sensationalism).
Recognize satire and misattribution: Examples include satirical sites and misrepresented figures (e.g., miscaptioned or AI-generated images).
Image forensics and AI-generated imagery: Common signs include symmetry anomalies, inconsistent hands/feet, unusual shadows, or other implausibilities; use AI-detection tools and critical inspection.
Fact-checking resources and practices:
Snopes highlighted as a trusted fact-check resource; not the only one, but a widely used one.
Use multiple fact-checking sources when in doubt and cross-reference primary sources when possible.
Understand the role of context: Provenance, date, author credentials, and publication standards affect reliability.
Practical guidance for verifying information:
When encountering unfamiliar sources, locate the byline or publisher and verify through independent background checks.
Be skeptical of sources with unclear ethics statements or without transparent accountability.
Use a stepwise approach: Stop, Investigate, Cross-check, and Trace to the original sources before citing.
Image and data literacy:
Learn to verify images and statistics with original datasets or credible statistical agencies (e.g., FBI crime data, official government/academic statistics portals).
Be cautious about selective data presentation and the way statistics are framed to support a narrative.
Research workflow and citation practices:
Use standardized citation styles as shorthand to facilitate quick identification of publication details (author, title, year, journal, volume, issue).
Tools and workflows: Research guides, “Ask a Librarian” services (text, chat, Zoom), and “Ask a Librarian” by text for quick questions.
Advanced search strategies and filtering: Use author, date, subject, and format filters to refine results efficiently.
Practical takeaways for students:
Leverage library resources to access paid content and avoid per-article charges.
Apply SIFT and other evaluation strategies to any information source encountered online.
Maintain transparency in citations and ensure attribution is clear and accurate.
When in doubt, consult librarians or research guides to validate sources and approaches.
Research Skills: Standards, Ethics, and Practice
Citation ethics and integrity:
Plagiarism concerns: Proper attribution of ideas and text; avoid presenting others’ work as your own.
The value of working through ideas and understanding content deeply—“you learn more by teaching others” as a mental model.
Navigating sources and expertise:
Trust in expertise should be contextual, not absolute; verify specialties and credentials when relying on experts.
Context matters: A known expert in one area may not be authoritative in another; verify the domain of expertise.
Tools for efficient research:
Research guides catalog databases and best sources for specific subjects.
Ask a Librarian services provide real-time help via text, chat, phone, or Zoom.
The library portal offers quick links to journals, periodicals, and subject-specific databases.
Practical Takeaways for Students: Assignments and Contact
Assignment logistics:
Three assignments planned; select a pair and locate a journal article from a reliable source; summarize the article or the underlying concept/theory.
Access and communication:
Jennifer’s email provided on Blackboard; students can email questions or arrange a meeting (Zoom, coffee) to discuss assignments.
Expectations:
Expect engagement: questions during and after the presentation; interactive class environment.
Summary of Key Concepts and Takeaways
GenAI basics and library relevance:
GenAI uses large datasets and probabilistic generation to respond to user prompts; data may be incomplete due to paywalls and data access restrictions.
Reliability and verification framework:
Use a structured approach (SIFT) and cross-check with primary sources; check for ethics statements and transparency; beware of satire and miscaptioned content.
Information literacy in practice:
Discovery vs. databases; use library resources to access paywalled content; practice incognito or privacy-aware searching when appropriate.
Critical thinking in AI-assisted work:
AI can streamline tasks but should not replace careful scholarly evaluation; always verify and cite sources appropriately.
Resources and support:
Clark Library guides; Snopes and other fact-checkers; FBI crime data; Stateline; library Ask-a-Librarian services; research guides for subject areas.
Final reminder:
Be conscientious about privacy, accuracy, and originality; use AI tool outputs to inform and not to substitute for rigorous scholarly work.
Appendices and References Mentioned
SIFT information literacy method: Stop, Investigate the source, Find better coverage, Trace claims to the original source; see Clark Library guide for details.
Snopes as a fact-checking resource for verifying claims and debunking misinformation.
Examples of image forensics cues for AI-generated images (e.g., hands, shadows, symmetry).
Data sources: FBI crime data and Stateline for statistics; caution about data quality and reporting gaps.
Research tools and services: Library discovery search, subject-specific databases, “Ask a Librarian” services, research guides, and advanced search options.
Contact information: Jennifer’s University of Michigan email (provided on Blackboard); instructor contact for questions and meeting arrangements.