Asking medical questions of AI with language spoken “in the wild”—meaning with LLM prompts from everyday consumers—brings back answers with decidedly mediocre accuracy, a new study shows. 

The devices were set aside to be scrapped before they were stolen. Now, patients could be using them and not even know it.

AMGA says it has the policy solutions ready and is urging Congress to act now before more practices are forced to make irreversible decisions.