Artificial intelligence is rapidly transforming weather prediction, enabling forecasts that once required hours of supercomputing time to run in just minutes. But as AI tools play an expanding role in high-stakes hazard modeling, researchers at Rice University say an essential question remains: Do AI-generated storms behave realistically? Their new study published in the Journal of Geophysical Research: Atmospheres provides a comprehensive evaluation of how AI-based global weather models simulate tropical cyclones.

Can artificial intelligence offer meaningful relationship advice, or do human therapists still provide something machines cannot replicate? Researchers from the Max Planck Institute for Psycholinguistics explored this question during a public experiment at the InScience Film Festival in the city of Nijmegen.

Augmented reality could transform job training for individuals with intellectual and developmental disabilities. In a new study, participants correctly completed just 14% of job-task steps on their own. With AR guidance, accuracy jumped to 93%, with some reaching 100%, and all achieved mastery across sessions. After only a 15-minute AR training session, participants reached at least 75% independence—progress that typically takes two to four months with traditional coaching—highlighting AR’s potential to accelerate workplace readiness and expand access to competitive employment.

A research team from the University of Tokyo and Tokyo University of Agriculture and Technology uncovered a new mechanism of Yaku’amide B, a deep-sea sponge-derived natural product. Using photoaffinity labeling, they found that yaku’amide B transiently binds CD9, inducing its degradation, in addition to inhibiting ATP synthase. This dual action suppresses cancer cell proliferation and migration, opening new avenues for anticancer drug development and protein degradation strategies.

Drug-drug interactions (DDI) can cause adverse drug reactions during the co-administration of multiple drugs, necessitating accurate and scalable prediction tools. While deep learning models have shown promise recently, most models show poor performance against drugs not encountered during training. Now, researchers have developed a lightweight and scalable model, called DDINet, designed specifically to predict unseen drug interactions. This innovative model achieves superior accuracy in predicting interactions for unseen drugs, with potential for practical deployment.