Archive · 6 Aug 2025
Google AI's Medical Mistake: Invented Brain Part Raises Alarm
Imagine a smart computer doctor that tries to diagnose illnesses.
Sometimes, it makes mistakes, like inventing a part of the brain that doesn't exist.
This happened with Google's Med-Gemini AI, which incorrectly identified a condition.
The AI mixed up real brain parts and blood vessels, creating a false diagnosis that was published in a medical paper.
Experts are worried because in medicine, it's very important to be accurate.
Even a small mistake can cause big problems for patients.
Google has fixed the error in a blog post but not in the original paper, causing more concern about the AI's accuracy.
Google's Med-Gemini AI created a nonexistent brain part in a medical diagnosis.
The AI misidentified the 'basilar ganglia' instead of 'basal ganglia'.
The error appeared in a published research paper despite review by Google's team, including medical experts.
Experts and practitioners express concerns about the AI's inaccuracies in healthcare.
Google has addressed the error in a blog post but not yet in the original research paper.
- Who
- Google's Med-Gemini AI.
- What
- Google's Med-Gemini AI created a nonexistent brain part in a medical diagnosis.
- When
- The error was identified in a research paper published in 2024.
- Where
- In a Google research paper.
- Why
- To identify medical conditions.
Concerns about AI Accuracy
Google's Perspective
Accuracy vs. 'Misspelling'
Concerns about AI Accuracy
Medical professionals emphasize the critical importance of accuracy in medical reports.
Google's Perspective
Google considers the mistake a 'misspelling'.
- AI Model
- Med-Gemini
- Error
- Incorrect brain part identification.
- Affected Term
- Basilar ganglia
- Intended Term
- Basal ganglia
- Expert Response
- Concerns about accuracy.
- Google's Response
- Silent fix in a blog post