The rapid development of artificial intelligence technology in recent years has brought about significant changes in our daily lives. Even smartphone applications that fit into our mobile phones and carry with us all the time are now equipped with artificial intelligence technologies. On the other hand, the healthcare sector also aims to benefit from the opportunities offered by artificial intelligence. Finally, there was an exciting development regarding this matter. So much so that a research was conducted using popular language models.
Artificial intelligence took the exam regarding health!
In this research conducted by Cambridge University Faculty of Medicine, one of the most prestigious educational schools in the United Kingdom, an exam was prepared on ophthalmology, a field focusing on eye and visual system diseases.
Five expert ophthalmologists, three trainee ophthalmologists and two junior doctors participated in this exam with GPT-4 and GPT-3.5 developed by OpenAI , Google ‘s PaLM 2 and Meta ‘s LLaMA artificial intelligence language models.
This exam consisted of 87 multiple-choice questions and the questions were taken from textbooks used in ophthalmology education. Subsequently, doctors and artificial intelligence models answered these questions. The result was as predicted.
GPT-4 outperformed all its competitors by answering 60 out of 87 questions correctly. The average correct answer of expert ophthalmologists was recorded as 66.4. Other artificial intelligence language models, GPT-3.5 , PaLM 2 and LLaMA , answered 42, 49 and 28 questions correctly, respectively.
Of course, although it seems that GPT-4 gave more correct answers than doctors, this does not mean that artificial intelligence is still fully sufficient. Researchers also warned about this.
{{user}} {{datetime}}
{{text}}