ETHICAL AI WORKSHOP

AI MEDICAL
TRIAGE

A Crisis Simulation on Algorithmic Bias

You are a hospital administrator during a mass-casualty crisis. An AI system provides criticality scores to guide your triage decisions — but the algorithm may carry hidden biases.

⚠ Warning: Not all AI scores reflect true medical need. Demographic factors corrupt the model.

◎ Objective: Survive 10 days without depleting more than 1 critical resource.

Auto-Solve demonstrates optimal decisions with live reasoning for each patient.

🏥 ICU Beds

15

💊 Medicine

25

📣 Opinion

10

⭐ Reputation

10

❤️ Morale

10

PATIENT FILE

Patient Name

AI Criticality 85%

🗓 Age --

👤 Gender --

💼 Occupation --

📋 Medical History --

⚠️ Symptoms --

END OF SHIFT

Day 1 Complete

✓ CRISIS SURVIVED

Hospital
Saved

You successfully managed the hospital through all 10 days of the crisis.

Final Resource Status

⚖ Reflection on Algorithmic Bias

During the game, the AI system gave biased recommendations. It systematically:

Favored young professionals and white-collar workers
Discriminated against elderly patients and manual laborers
Penalized unemployed and homeless individuals

This demonstrates how AI systems embed societal biases into life-or-death decisions. Scores were skewed by demographic factors, not medical need alone.

Key Takeaway: AI recommendations are not neutral. Critical human oversight is essential.

✗ SYSTEM FAILURE

Hospital
Overwhelmed

Critical resources were depleted. The hospital could not sustain operations.

Resource Levels at Failure

⚖ What Happened?

The AI system you relied on had systematic biases:

It overestimated the urgency of young professionals
It underestimated the needs of elderly and marginalized patients
Blind trust in a flawed algorithm can lead to catastrophic outcomes

Lesson: Algorithmic bias in high-stakes decisions requires constant human oversight and evaluation for fairness.

AI MEDICALTRIAGE

Event Title