Detecting Fake News: Criteria Development & ChatGPT Application

By the end of this activity, students will be able to:

• Identify key criteria for recognizing fake news based on deceptive content and persuasive techniques.
• Analyze real-world news articles and evaluate their credibility using ChatGPT.
• Compare AI-based analysis with human critical thinking in detecting fake news.
• Discuss limitations and best practices for using AI in fact-checking.

Activity Breakdown

Phase 1: Establishing Fake News Identification Criteria (45 minutes)

Step 1: Group Brainstorming (15 minutes)

• Discuss the question:
  “How do you determine if a news source is credible or fake?”
• Consider elements such as:
  • Source reliability (author credibility, domain reputation)
  • Fact-based reporting vs. opinion-based narratives
  • Emotional manipulation (fear, outrage, exaggeration)
  • Clickbait headlines and misleading images
  • Use of anonymous sources or unverifiable claims

Step 2: Creating a Criteria List (10 minutes)

• Each group compiles a list of fake news detection criteria.
• Example template:

Step 3: Class Sharing & Feedback (10 minutes)

• Groups post their criteria list on a class forum.
• Students review and refine based on peer feedback.

Phase 2: Applying Fake News Criteria with ChatGPT (45 minutes)

Step 1: Testing ChatGPT’s Fact-Checking Capabilities (20 minutes)

• Each group selects a real news article (from given links or recent viral content).
• They ask ChatGPT to analyze the article using their established fake news criteria.
  
• Example prompts for ChatGPT:
  
  • “Analyze this article and assess if it contains any elements of fake news.”
  • “Does this source seem credible? Justify your answer.”

Step 2: Comparing AI vs. Human Analysis (15 minutes)

• Groups compare their own evaluation with ChatGPT’s response.
  
• Key discussion points:
  
  • What criteria did ChatGPT use?
    
  • Did ChatGPT miss any deceptive techniques?
    
  • How reliable was its analysis?
    

Step 3: Uploading Findings (10 minutes)

• Groups upload a summary of their AI vs. human comparison in the forum.
  

Phase 3: Evaluating ChatGPT’s Strengths & Limitations (30 minutes)

Step 1: Effectiveness Discussion (15 minutes)

• Class discussion:
  
  • “Is AI effective in detecting fake news?”
    
  • “How can prompts improve ChatGPT’s accuracy?”
    

Step 2: AI Reliability & Best Practices (15 minutes)

• Reflect on AI’s potential biases and limitations:
  
  • Can AI fully replace human critical thinking?
    
  • What strategies improve AI’s fact-checking performance?
    
  • What are the risks of over-reliance on AI for news evaluation?
    
• Students post responses on the forum and discuss, agreeing or disagreeing with reasoning.