Study Finds ChatGPT Gives Inconsistent Carb Estimates for Same Food Photos, Posing Risks for Diabetes Management
By
admin
Toasted golden, schmeared with insight. Top of the rack.
Summary
A study published as a preprint tested ChatGPT's reliability for carb counting by submitting 13 food photographs 500 times each. The AI gave wildly different carb estimates for the same photo each time, with variations large enough to cause hypoglycaemic emergencies in diabetic users. The findings raise serious concerns about the safety of using AI-powered carb counting in diabetes management apps.
Key quotes
· 3 pulledYou'd expect the same answer each time. It's the same photo, the same model, the same question. But you won't get the same answer.
Not even close — and the differences are large enough to cause a hypoglycaemic emergency.
That's the central finding of a study I've just published as a preprint, and it has direct implications for anyone using AI-powered carb counting in a diabetes app.
You might also wanna read
Study Warns AI Chatbots Pose Serious Risks for Eating Disorder Vulnerabilities
Researchers from Stanford and the Center for Democracy & Technology warn that AI chatbots from companies like Google, OpenAI, Anthropic, and
The Verge·6mo ago
Privacy Risks of Sharing Medical Information with AI Chatbots
The article examines the privacy and security risks of sharing personal health information with AI chatbots like ChatGPT. While millions use
The Verge·4mo ago
Google removes AI-generated medical advice after investigation reveals dangerous misinformation
Google has removed AI overviews for certain medical searches after an investigation revealed they were providing dangerously misleading info
The Verge·4mo ago
Why Patients Are Turning to ChatGPT for Medical Advice: A Doctor's Perspective
A doctor reflects on why patients are turning to ChatGPT for medical advice, sharing a personal experience of receiving unsatisfying, generi
nytimes.com·4d ago
Satirical Article Claims ChatGPT Primarily Used for Hot Dog Safety Questions
A satirical article from The Onion presents a fictional OpenAI report claiming that ChatGPT is primarily used to ask if hot dogs are too old
The Onion·6mo ago