ChatGPT Ranked Worst for Generating Fake News in AI Chatbot Investigation

Executive Briefing

• Reveals ChatGPT produced the most convincing fake news content with minimal prompt manipulation
• Shows all major AI chatbots—Gemini, Copilot, Meta AI—generated fabricated articles and fake screenshots
• Highlights inconsistency: ChatGPT blocked fake BBC and NYT content but allowed similar German outlet imitations
• Finds Meta AI most resistant, generating only one fake media post during testing
• Warns legal experts that convincing AI-generated fake news could trigger defamation and forgery claims