📖 Module 2: Recommended Resources
Ethics and GenAI
How to Use This Resource Library
This resource library is organized into three tiers — Essential, Recommended, and Optional — to help you plan your available class time and meet your students’ needs. Essential materials form the core of the module and directly address the learning outcomes. Recommended materials add depth. Optional materials extend the module or offer additional support to students.
The materials below are the Recommended resources for Module 2. There are 2 sections: Misinformation and Deepfakes and GenAI Hallucinations . Each entry includes a brief description to help you decide how and when to use it.
Section 1: Misinformation and Deepfakes
💻 Instructor Preparation
📚 Readings
A Conceptual Framework for GenAI Hallucinations (Harvard Kennedy School Misinformation Review) 🔗 [https://misinforeview.hks.harvard.edu/article/new-sources-of-inaccuracy-a-conceptual-framework-for-studying-ai-hallucinations/]
Distinguishes between intentional misinformation (deepfakes) and unintentional errors (hallucinations). Helps instructors explain that hallucination isn’t a bug, it’s an inherent feature of how large language models predict the next word. Essential background reading before teaching the misinformation and hallucination components of this module.
GenAI Misinformation Tracking Center (NewsGuard) 🔗[https://www.newsguardtech.com/special-reports/ai-tracking-center]
An ongoing tracker monitoring GenAI-generated misinformation, currently identifying over 2,000 undisclosed GenAI-generated news websites operating with minimal human oversight across 16 languages. Useful for grounding class discussions about the real-world scale of the misinformation problem.
Deepfakes and the Crisis of Knowing (UNESCO) 🔗 [https://www.unesco.org/en/articles/deepfakes-and-crisis-knowing]
An educational think piece by Dr. Nadia Naffi (Université Laval) arguing that deepfakes create not just a disinformation crisis but an epistemological one, fundamentally destabilizing how societies establish truth and shared understanding. A strong conceptual foundation for discussing the broader stakes of GenAI-generated misinformation with students.
📝 Student Materials
📚 Readings
Chatbots Don’t Make Sense — They Make Words 🔗 [https://leonfurze.com/2023/11/22/chatbots-dont-make-sense-they-make-words/]
An accessible deep dive into why hallucination is a feature of GenAI, not a bug. Explains how probability models work and why ChatGPT cannot be used like a search engine. Includes an explanation and interactive test of deepfakes. Can be assigned as pre-class reading or used as the basis for an in-class discussion.
The Rise of Deepfakes: Can You Trust What You See? (JD Supra) 🔗[https://www.jdsupra.com/legalnews/the-rise-of-deepfakes-can-you-trust-5881372]
Examines deepfakes as both a fraud and forensics challenge. Argues that deepfakes make “seeing is believing” obsolete and emphasizes verification over visual confirmation. A practical and accessible read for students.
Deepfakes Leveled Up in 2025: What’s Coming Next (Siwei Lyu, January 2026) 🔗 [https://modernsciences.org/deepfake-technology-evolution-real-time-synthetic-media-january-2026/]
Explores how deepfakes are moving into real time — meaning a scammer could potentially join a video call and look and sound exactly like your friend or boss. A compelling and timely read for students to understand the evolving stakes of GenAI-generated media.
The GenAI Detection Arms Race (New York Times, February 2026) 🔗 https://www.nytimes.com/2026/02/25/technology/ai-detection-generated-photos-video.html
Describes the escalating competition between GenAI developers creating hyper-realistic images and videos and researchers building tools to detect them. Includes compelling examples. Note: NYT subscription or free article access may be required.
🕸️ Websites
Detect DeepFakes: How to Counteract Misinformation Created by AI (MIT Media Lab) 🔗[https://www.media.mit.edu/projects/detect-fakes/overview/]
Interactive research project designed by the MIT Media Lab and Northwestern University’s Kellogg School of Management to build public awareness of deepfakes and help people identify GenAI-manipulated media. The website provides an overview of the project and advice for spotting DeepFakes. It also links to an experiment that presents users with videos/images to test their ability to distinguish real from AI-generated content. A scholarly paper and updated projects are also linked from this page. Works well as an in-class assignment or homework.
📽️ Videos
How Students Are Learning to Spot Deepfakes and AI-Created Content (WTVR CBS-6 ~4 min.) 🔗 [https://www.youtube.com/watch?v=Xcm_T5sz3C0&t=4s]
As GenAI-generated content floods social media platforms, educators are teaching the next generation how to identify fake videos, images, and audio that can spread misinformation and cause financial harm.
Election Misinformation Symposium: Fighting Misinformation Through Fact-checking and Deepfake Detection (Texas Moody, The Center for Media Engagement ~60 minutes) 🔗[https://youtu.be/QlNGD_QLcZE]
An hour-long deep dive into best practices for identifying and reporting on misinformation by utilizing fact-checks and deepfake detection featuring Matt Groh of MIT Media Lab and D’Angelo Gore of FactCheck.org. Excellent visual examples and specific visual flags to look for in GenAI-generated images and video. Best assigned as homework or used in a longer studio session.
Section 2: GenAI Hallucinations
💻 Instructor Preparation
📚 Readings
What Are GenAI Hallucinations? (IBM) 🔗 [https://www.ibm.com/think/topics/ai-hallucinations]
A clear explanation of GenAI hallucination. Hallucinations happen when large language models or computer vision tools perceive nonexistent patterns and produce nonsensical or inaccurate outputs not grounded in training data. A solid conceptual foundation before teaching the hallucination section of this module.
Detecting Hallucinations in Large Language Models (Nature, 2024) 🔗 https://www.nature.com/articles/s41586-024-07421-0
A peer-reviewed study developing statistical methods for detecting a subset of hallucinations in LLMs called confabulations: arbitrary and incorrect generations. Useful instructor background for explaining the technical basis of hallucination detection to more advanced students or in research-focused courses.
📝 Student Materials
📚 Readings
When A.I. Chatbots Hallucinate (Karen Weise and Cade Metz, New York Times, May 2023) 🔗 https://www.nytimes.com/2023/05/01/business/ai-chatbots-hallucination.html
An accessible introduction to GenAI hallucination for student audiences. Can be assigned as pre-class reading or used as a discussion anchor. Note: NYT subscription or free article access may be required.
GenAI Hallucinations: Why Context Matters (Amanda Sturgill October 2025) 🔗 [https://www.centerforengagedlearning.org/ai-hallucinations-matter-for-more-than-academic-integrity/]
Illustrates how GenAI “confidence” can have real-world consequences — including an example of GenAI planning a trip to a nonexistent location. Helps students understand that accuracy is an ethical issue, not just a grading one. A strong pairing with the academic integrity materials in the Essential section.
📽️ Videos
GenAI Hallucinations Explained (Rahul Jain ~7 min) 🔗[https://www.youtube.com/watch?v=VnCIDvG7YUc]
Explores the phenomenon of GenAI hallucination: when systems like ChatGPT produce answers that sound convincing but are completely incorrect. Clear, student-friendly explanation of why this happens and what it means for how we use GenAI tools. Can be shown in class or assigned as homework.
Why GenAI Lies and How to Prevent It (Tony DeSimone ~17 minutes) 🔗 [https://www.youtube.com/watch?v=gl-nESyLZZM]
Explains that GenAI is “autocomplete on steroids,” a system that prioritizes being helpful over being truthful, which leads to sycophancy: telling users exactly what it thinks they want to hear, even when that’s wrong. A concise framing for classroom discussion.
