A study led by Dr. Jonathan H. Chen and his team explores how AI-powered chatbots can support doctors in making nuanced clinical decisions. While chatbots have shown proficiency in diagnosing diseases, their effectiveness in guiding treatment and care decisions (like when to stop blood thinners before surgery or adjusting treatment protocols based on past adverse reactions) has been less clear.
The research found that chatbots outperformed doctors who only had access to internet searches and medical references. However, when doctors used their own LLM (large language model), they performed just as well as the chatbots. This suggests that a combination of human judgment and AI can lead to better decision-making than either alone.
The study, published in Nature Medicine, highlights how AI can assist in clinical management reasoning, where decisions aren’t always black and white, and depend on context rather than textbook rules. This research prompts a reevaluation of how AI and humans should work together in healthcare.