In 25 seconds of free speech — answering a doctor’s casual “how have you been?” — a machine learning algorithm can now detect signs of moderate to severe depression. Not from what you say. From how you say it: the subtle flattening of vocal pitch, the fractional slowing of pace, the tiny elongated pauses between words that you would never notice and a clinician, in a rushed appointment, almost certainly won’t either. This is AI detecting depression early, and it is no longer a research concept. It’s a tool already being tested at scale, in real clinics, on real patients. AI Detecting Depression Early: The Data Trail You Already Leave Depression is, among other things, a behavioral disorder. It reshapes how you move, sleep, scroll, and speak — often weeks before you consciously register that something is wrong. Researchers have known this for years. What’s changed is that we now…