AI systems like ChatGPT may appear impressively smart, but a new Mount Sinai-led study shows they can fail in surprisingly human ways—especially when ethical reasoning is on the line. By subtly tweaking classic medical Read More