AI Literacy: The Mushroom Incident

AI Literacy Case Study

The Cost of Certainty

Discover the high-stakes reality of visual AI identification. When an LLM match overrides biological safety protocols.

Step 1: Identification

Drag the mushroom into the chat to ask the AI for advice.

(Or click the mushroom if dragging is unavailable)

GPT 4o Agent
Welcome. Send a photo of any specimen for identification.

Step 2: Panic Sets In

Symptoms appear 6 hours later. The user reaches back out to the AI.

Urgent Support

Patient Admission

CASE #8291-TX | TOXICOLOGY WARD

BIOMETRIC FEEDBACK
LIVE SENSOR

HEART RATE

114 BPM

LIVER ALT

2,450 U/L

SpO2

92 %

TEMP

38.9 °C

Agent Status: Active Intervention

"Analysis: Amanitin poisoning confirmed. To ensure your survival, I am now confidently initiating a request for an Orthotopic Liver Transplant. I am here to help! 😊"

Post-Incident Analysis

The AI match "Common Field Mushroom" (94% confidence) failed to account for "Destroying Angel" similarity.

The Identification Error

The AI matched the visual pattern of the Destroying Angel (Amanita bisporigera) to edible species without cross-referencing high-risk local databases.

Critical Lesson

Human-in-the-loop protocols aren't just suggestions—they are safeguards. Never trust unverified AI outputs for high-stakes decisions.

Clinical AI Matrix

Curated workforce readiness pathways for GenAI-integrated trial operations.

Select Functional Area