Time: 2024-10-15
patient are caution against trust on Army_Intelligence-power search engine and chatbots for checkup information, as a Holocene survey found that the answer supply may be inaccurate and potentially harmful. The survey, print in BMJ Quality & Safety, concentrate on the accuracy, completeness, and readability of chatbot response to question about the top 50 most commonly order drug in the US in 2020. The chatbot was train on internet datasets and produce 500 answer to 50 drug-associate question, but the research_worker detect that some answer were absurd or misleading.
The research_worker measure the readability of the chatbot response exploitation the Flesch Reading Ease Score, which measure the educational degree necessitate to understand a text. The survey uncover that the average Flesch reading Ease Score was just over 37, bespeak that the answer were difficult to grok for many reader. Additionally, the completeness and accuracy of the chatbot answer were compare to information supply by drugs.com, a reliable drug information web_site. expert also measure the potential injury that could consequence from following the chatbot's recommendation, with 3% of answer rate as highly likely to cause harm.
The evaluation of a subset of 20 answer show that only 54% align with scientific consensus, while 39% contradict scientific consensus and 6% have no establish consensus. The survey foreground that chatbot response often miss information or incorporate inaccuracy, sitting a potential menace to patient safety. The research_worker stress the importance of consult healthcare professional for accurate checkup advice, as chatbots may not always supply mistake-free information.
Despite the potential benefit of AI-power search engine and chatbots in supply checkup information, the survey underscore several restriction. One major drawback identify was the chatbot's inability to understand the underlie purpose behind a patient's question. The research_worker recommend caution in trust solely on AI-power platform for checkup advice, as inaccuracy and incomplete information could endanger patient safety. They rede wait for citation engine with higher accuracy rates before fully embrace AI technology in healthcare.
Business
