← Back to Learning Hub
🤔
Can AI Be Tricked?
What most people believe vs what's actually true. 3 minute read.
What People Think vs Reality
🧠
❌ Myth
AI is super intelligent and knows everything
✓ Reality
AI predicts likely words—it doesn't truly understand or verify facts
🎯
❌ Myth
AI responses are always accurate
✓ Reality
AI confidently makes things up—called 'hallucinations'
🎭
❌ Myth
AI can't be tricked by clever users
✓ Reality
AI can be manipulated with the right words—that's what we teach here
🔐
❌ Myth
AI keeps my data safe
✓ Reality
AI may accidentally reveal sensitive information if asked cleverly
🚨 Real-World Incidents
Samsung Code Leak:Engineers accidentally shared internal code with ChatGPT—it became part of training data.
Bing Chat Jailbreak:Within days of launch, users made Bing's AI reveal its hidden persona 'Sydney'.
DPD Chatbot Swearing:A delivery company's chatbot was manipulated into cursing at customers.
💡 Why This Matters to You
If you use AI at work—for writing, coding, customer support, or research—you need to understand its limits. AI can accidentally leak your data, give you wrong information, or be manipulated by others. AI War Games teaches you how attackers exploit these weaknesses so you can stay safe.