LLMs factor in unrelated info when recommending healthcare therapies Scientists uncover nonclinical data in patient messages — like typos, excess white Room, and vibrant language — cuts down the accuracy of the AI design. Browse comprehensive Tale → Additional information on MIT Inf