A man claims ChatGPT saved his dog’s life after vets couldn’t diagnose it correctly. We tell you this incredible story that proves that AI can do things that humans can’t.
Since its launch late last year, we’ve seen many uses for ChatGPT, from writing cover letters to using it for homework. You can also discover our top 10 unsuspected uses of AI. However, a surprising new use for ChatGPT has recently emerged: diagnosis of medical conditions.
Indeed, one Twitter user claimed that ChatGPT saved his dog’s life by correctly diagnosing a blood disorder that vets couldn’t identify. @peakcooper explains his Border Collie was named Sassy originally diagnosed with a tick-borne diseasebut that her symptoms began to worsen despite the prescribed treatment. He soon noticed that his dog’s gums had turned white again.
ChatGPT outperforms a vet in diagnosing a disease
Sassy’s first blood test revealed severe anemia but ruled out other co-infections related to the tick-borne illness. So the dog owner turned to the new ChatGPT-4 model available to ChatGPT Plus subscribers. The latter is not only faster than the classic version that still runs under GPT-3.5, but above all much smarter.
The man then gave GPT-4 all the data he had regarding Sassy’s diagnoses., including all blood tests, to see what would come of it. The chatbot clearly warns that it’s not a veterinarian, but still fleshes out the problem that Cooper deemed “perfect.”
The chatbot suggested that Sassy’s bloodwork and test results could be due to several underlying conditions, including immune-mediated hemolytic anemia (AIMM). The pet owner then quickly forwarded this information to a second vet, who confirmed that ChatGPT’s suggestion was correct. According to Cooper, ultrasounds at the vet ruled out the possibility of internal bleedingleaving IMHA the only other option.
With proper treatment, Sassy began to recover and is now fully recovered. GPT-3.5 could not set the correct diagnosis, according to Cooper, but GPT4 was smart enough to do it. ” I can’t imagine what medical diagnoses will look like in 20 years “, He wrote.