ChatGPT to show its potential for ever-growing medical knowledge and experience is demonstrated by the researchers at the University of Kansas Medical Center. The research team has stated that globally clinical knowledge is expanding and the requirement of new medical literature and practice guidelines has increased.
Therefore, the study reveals that the articles are added frequently; to guide literature and many more relevant clinical reviews featuring in article abstracts, however, reviewing articles carrying up to 300 words to make it a bit time-consuming.
Furthermore, the recent development in AI technology is investigating whether a large language model (LLM) could be used by clinicians to systematically review medical literature. The team selected ChatGPT-3.5 tool tasks to summarise 140 peer-reviewed abstracts from 14 journals. However, to increase LLM’s performance human physicians test the accuracy of ChatGPT-generated summaries. Therefore the researchers compared the rate of ChatGPT and clinicians in medical specialities through abstracts.
LLM was found to provide summaries that are 70% shorter than the actual abstract and it did not change the actual meaning of the original abstract. Also, a more relevant summary that is classified by the clinicians is similar to the performance of LLM. However, ChatGPT does generate accurate summaries, but somehow it was less successful in identifying the original relevance.
The researchers have compared the rating of ChatGPT with human physicians rates, and the result shows that the ChatGPT 3.5 model is not quite ready to do that yet. It works well in identifying the primary and journal sources more accurately, but it is not so good in identifying articles related to medical findings.
Moreover, the model of ChatGPT is designed to detect healthcare conditions and help physicians, but in some critical medical cases, it still lacks evaluations to be made with more research. The researchers also indicate the new version of ChatGPT that is going to be released soon will come with better relevance in determining medical conditions more accurately.