This post is also available in: heעברית (Hebrew)

ScholarAI is a new plugin for ChatGPT that lets users search for academic literature without receiving hallucination responses.

Despite being used mostly for text production and language translation, interest in ChatGPT’s potential uses in health care is growing. ChatGPT is spearheading a paradigm shift in the realm of cardiothoracic surgery research, through the joint use of artificial intelligence (AI) and natural language processing (NLP). The potential gains from the possible integration of ChatGPT within cardiothoracic surgery research are promising.

According to Interesting Engineering, ChatGPT harnesses the power of machine learning through the use of a strong language model that was trained on a variety of data sources like academic literature and medical textbooks. The chatbot’s advantage and strength is its ability to examine and comprehend vast amounts of data, then provide significant insights that could revolutionize cardiothoracic research and practice.

Nevertheless, ChatGPT has a serious hallucination problem that is being worked on by many experts but is not yet solved.

To deal with this issue a team around Duke University developed the ScholarAI plugin, which provides better results than ChatGPT itself and provides real papers from respected journals. The monetized version “ScholarAI Premium” can even provide tables and figures from papers, though this is still in beta so results may vary from paper to paper. The option “Paper Question Answering” can search a PDF to find the best parts of the paper to answer a given question.

Now researchers and clinicians will be able to use this new tool to better treat patients, and the world of medical research and practice is facing immense improvement and future advancements.