AI mistakes can cost doctors time when writing to patients
Reports and Proceedings
Updates every hour. Last Updated: 29-Oct-2026 01:16 ET (29-Oct-2026 05:16 GMT/UTC)
A Dartmouth study finds that AI may cost doctors time when corresponding with patients by making them spend more time editing AI-drafted responses than it would take to write them. The researchers analyzed 146,000 conversations between patients and their primary care physicians from the online portal of a large healthcare system and found that automated responses frequently introduced errors and extraneous details. But the study provides a framework for generating "good" AI responses that align with how clinicians communicate and shows that adapting AI to individual physicians can reduce editing by 26%. The study was presented at the 64th Annual Meeting of the Association for Computational Linguistics.
New research from Northwestern University has found, like humans, mice also can take a single sniff to deliberately probe their environment — something scientists previously did not know.
Two new complementary Northwestern University studies, which will be published together July 3 in Science Advances, studied olfaction from opposite ends and found rodents and humans rely on the same underlying neurophysiology — the brain’s motor and rhythmic building blocks — to process smells. Although a mouse’s single sniff is much shorter than a human’s, the underlying tempo of smell processing is the same, according to the findings. The results suggest these sensory systems are fundamentally similar and have been preserved through evolution.
Taken together, the findings from the two labs suggest something important: Mammals all rely on a similar underlying system for smell, with each species putting its own twist on the same basic design. The work answers a fundamental question: How do mice and humans sample our environment so we understand it and predict the next thing we want to do?
Genomic surveillance — the process of monitoring and sequencing pathogens — is one of the most important tools for detecting emerging viral threats. But global surveillance systems remain costly, unevenly distributed and often are too slow to identify dangerous variants before they spread internationally, amplifying future threats of disease outbreaks.
A recently published research paper in Nature Communications, co-authored by Dr. Patricia Ning, assistant statistics professor, and Jifan Li, a doctoral candidate in the Department of Statistics, along with collaborators from multiple international institutions, introduces a new framework to rectify these issues while using fewer resources, making genomic surveillance rapid and cost-effective in preparation for new strains of COVID-19.