At a time when companies are embarking on a frantic race for artificial intelligence (AI), the ChatGPT-4.0 large language model (LLM) continues to progress. This is evidenced by its resounding success in clinical neurology, which suggests possible practical applications.
OpenAI’s ChatGPT-4.0 AI validates clinical neurology test at 85%
ChatGPT-4.0, OpenAI’s cutting-edge language model, is once again in the news. The AI tool achieved a remarkable feat by passing a neurology clinical exam. And talk about“feat” is not an emphasis.
In fact, the ChatGPT-4.0 AI model provided 85% true answers to questions asked of it. In particular, in the context ofa proof-of-concept study carried out by researchers from Heidelberg University Hospital and the German Cancer Research Center.
To test ChatGPT-4.0, researchers relied on questions from the American Board of Psychiatry and Neurology. It is an American organization dedicated to the evaluation and certification of physicians specializing in psychiatry and neurology.
The experiment included comparing ChatGPT-4.0 to its predecessor, ChatGPT-3.5. And the results are incommensurable. Where ChatGPT-4.0 achieves a success rate of 85%, ChatGPT-3.5 does no better than a score of 66.8%. This shows the considerable AI progress made by the latest iteration.
A development potentially favorable to clinical applications
One of the high points of this study is to have highlighted the excellent reaction of these AI models. Particularly on behavioral, cognitive and psychological issues.
However, they show weaknesses when it comes to tasks requiring high-level thinking. Despite this, the overall performance of the ChatGPT-4.0 AI model exceeds the average human score.
This breakthrough suggests promising future applications in clinical neurology. An option that should only be seriously considered once the necessary improvements have been made.
Researchers recognize the potential of LLMs in healthcare documentation and decision support systems. They still invite neurologists to approach their practical use with the greatest caution. This is due to their weakness in tasks involving high-level cognitive functions.
Maximize your Tremplin.io experience with our ‘Read to Earn’ program! For every article you read, earn points and access exclusive rewards. Sign up now and start earning benefits.
