How AI support can go wrong in safety-critical settings
Peer-Reviewed Publication
Updates every hour. Last Updated: 18-Nov-2025 17:11 ET (18-Nov-2025 22:11 GMT/UTC)
When it comes to adopting artificial intelligence in high-stakes settings like hospitals and airplanes, good AI performance and a brief worker training on the technology is not sufficient to ensure systems will run smoothly and patients and passengers will be safe, a new study suggests. Instead, algorithms and the people who use them in the most safety-critical organizations must be evaluated simultaneously to get an accurate view of AI’s effects on human decision making, researchers say. The team also contends these evaluations should assess how people respond to good, mediocre and poor technology performance to put the AI-human interaction to a meaningful test – and to expose the level of risk linked to mistakes.
Researchers have identified key factors that influence the flavour of chocolate during the cocoa bean fermentation process, a discovery that could offer chocolate producers a powerful tool to craft consistently high-quality, flavour-rich chocolate.
Researchers have developed a self-powered microneedle patch to monitor a range of health biomarkers without drawing blood or relying on batteries or external devices. In proof-of-concept testing with synthetic skin, the researchers demonstrated that the patches could collect biomarker samples over periods ranging from 15 minutes to 24 hours.
Farmers in Arkansas, Louisiana and Mississippi, pay four to six times more for crop insurance than their counterparts in the upper Midwest, and Hunter Biram wanted to know why. The result of his research with colleagues in Kansas was published in late July in the American Journal of Agricultural Economics.