The veterinarian cannot detect the disease of his dog, ChatGPT succeeds and saves her

Artificial intelligence is increasingly used in the field of health, particularly in aiding the diagnosis and early detection of diseases. While these are currently mainly experimental solutions dedicated to professionals, an Internet user reportedly recently used GPT-4, the latest version of ChatGPT, the OpenAI chatbot, to save the life of his dog, reports Numerama.

A medical wandering

In his testimony published on March 25 on Twitter, this Internet user named Peakcooper first explained that his dog Sassy had fallen ill after being infected with ticks. The man then went to a veterinary clinic, where treatment was administered to his animal. His condition seemed to improve at first but eventually got worse.

After analysis, the veterinarians noticed that his hemoglobin level in the blood had still dropped abnormally. If the doctor had suspected a co-infection, the tests for it eventually came back negative. He then invited the owner to “wait and see what happens” next.

Tracked by ChatGPT

However, seeing his dog’s condition deteriorate, Peakcooper did not want to remain inactive. After making an appointment with a second vet, he had the idea to ask ChatGPT the question. “It occurred to me that medical diagnostics seemed like the kind of thing GPT4 could potentially be very good at,” he recounted. To put the chatbot on the way, he gave her the results of his blood tests.

If ChatGPT immediately told him that he was not a veterinarian, he nevertheless presented him with several lines of thought, including the hypothesis of an immune-mediated hemolytic anemia (IMHA). With further tests and the opinion of the second vet, this diagnosis was finally confirmed. “We put the dog on treatment and she almost completely recovered,” he said. AHMI is a common affliction, however, and would have been the most likely here, leading the owner today to wonder why the first vet didn’t catch it.

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.