OpenAI’s Artificial Intelligence Model Overtakes Ophthalmologists

When OpenAI is mentioned, the first thing that comes to mind is ChatGPT-4 and text or image creation. However, it should not be forgotten that the GPT-4 model is actually an advanced model that can be used in different areas. OpenAI The latest artificial intelligence model developed by A.Ş. reveals how advanced new systems can be in the field of healthcare.

According to an article published Wednesday GPT-4 modeli generally achieved a higher success rate than doctors in terms of diagnosing eye diseases and providing treatment recommendations. Even at points he could not reach, he achieved success rates very close to doctors. This success rate was also an important development for artificial intelligence, especially considering that artificial intelligence sometimes “hallucinates” and uses data that does not exist.

Left non-specialized doctors behind

In the study, GPT-4 competed with expert ophthalmologists, doctors with specialist training, and doctors without any specialization. In studies using 87 patient scenarios, GPT-4 not only outperformed non-specialist doctors also with experts It got nearly head-to-head results.

This study gained attention because it compared GPT-4 against healthcare professionals, not outcomes. In addition, in this study, generative artificial intelligence was used with its general structure, that is, it works in a narrowed area such as cancer tissue identification examples. artificial intelligence not used.

Artificial intelligence models are being used more and more every day, especially in the diagnosis of diseases such as cancer, and our computers are preferred to make calculations instead of doctors or scientists. Thus, new breakthroughs are opened. On the other hand, these innovations bring the need for good data and the need for data auditing. brings. Additionally, when we have a health problem, we need to consult a healthcare institution, not Google or ChatGPT.