A recent study reveals that leading AI models like GPT-4o can be easily manipulated into creating convincing and dangerous health disinformation, exposing critical vulnerabilities in their current safety features.
Most Recent Articles
Distinct Transcriptional Pathways Regulate Inflammatory and Lysosomal Responses in Particle-Activated Macrophages
New research reveals that macrophage responses to pathogenic particles are governed by two independent signaling pathways. This finding could unlock novel therapies for inflammatory diseases like gout and silicosis.
CSF1R+ LAMM Cells Identified as a Key Mediator of CAR-T Therapy Resistance in B-Cell Lymphoma
New research uncovers how specific immune cells sabotage CAR-T therapy in B-cell lymphoma. Preclinical results show that a combination treatment could significantly improve patient survival by overcoming this resistance.
Mendelian Randomization Links Higher DHEA-s to Reduced Lifespan in Men
A new genetic study challenges DHEA's reputation as an anti-aging supplement. It finds that higher levels of this supplement are associated with a significantly shorter lifespan and increased blood pressure in men.
A Validated Model for Estimating Arm SBP from Ankle Measurements and Its Prognostic Value
A new predictive model offers a more accurate way to interpret ankle blood pressure, addressing a critical gap in hypertension care for patients unable to undergo standard arm measurements.