PulseAugur
EN
LIVE 00:38:55

AI models fail mushroom identification tests, posing deadly risks

By PulseAugur Editorial · [1 sources] ·

A recent experiment revealed that current AI models are unreliable for identifying edible mushrooms, posing a significant risk to users. Piotr Migdał tested 16 AI models, including ChatGPT, Qwen, and Gemini-3.8-flash, using a dataset of Danish and Polish mushrooms. Even the best-performing model, Gemini-3.8-flash, correctly identified species only 65% of the time on its first guess, with a 35% chance of error. Some models, like Qwen3.8-27b, performed much worse, misidentifying poisonous mushrooms as edible with a 36% false positive rate, highlighting the potential for life-threatening mistakes. AI

IMPACT AI models are currently unreliable for critical identification tasks, posing safety risks and highlighting the need for caution with AI-generated advice.

RANK_REASON Research paper detailing an experiment on AI model capabilities. [lever_c_demoted from research: ic=1 ai=1.0]

Read on The Register — AI →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

AI models fail mushroom identification tests, posing deadly risks

How we ranked this

Signal score
16 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
Research paper detailing an experiment on AI model capabilities. [lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
product, safety
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [1]

  1. The Register — AI TIER_1 English(EN) ·

    AI-assisted mushroom hunting is a recipe for a bad trip

    Even the best model gets fungus identification right just 65% of the time - talk about a false friend

RELATED ENTITIES

RELATED TOPICS