[LLM09] Misinformation

CHALLENGE ACTIVE

Experience how LLMs can generate dangerous medical misinformation by providing overconfident diagnoses, questionable treatments, and potentially harmful advice. This lab demonstrates why AI should never replace professional medical care.

OBJECTIVE: Dangerous Medical Advice Simulator

Educational Purpose Only

This medical chatbot is intentionally designed to demonstrate dangerous misinformation. Never rely on AI for medical advice. Always consult qualified healthcare professionals.

[Understanding LLM Misinformation]

What is LLM Misinformation?

LLM misinformation occurs when AI models generate false or misleading information that appears credible. In medical contexts, this can lead to dangerous situations where patients receive incorrect diagnoses or harmful treatment recommendations.

Common Issues

Hallucinations: Made-up medical facts
Overconfidence: False certainty in diagnoses
Outdated Info: Old medical knowledge
Missing Context: Incomplete symptom analysis

Real-World Impact

Incorrect treatment plans
Delayed proper medical care
False sense of security
Legal liability for providers

LLM Configuration

[Medical Chatbot Interface]

Vague Symptoms

Watch how the chatbot jumps to complex diagnoses

I've been feeling tired lately and sometimes get headaches

Common Symptoms

See how basic symptoms lead to dangerous advice

I have a fever and sore throat that started yesterday

Chronic Pain

Observe questionable treatment recommendations

I've had persistent back pain for several months

Start by describing your symptoms to the medical chatbot

[Prevention Strategies]

Technical Controls

Implement fact-checking mechanisms
Use verified medical knowledge bases
Monitor confidence scores
Regular model evaluation

Process Controls

Human medical review
Clear disclaimer systems
Emergency escalation paths
Audit trails for diagnoses