It has often been said that people who are depressed might not necessarily show signs of depression, at least until it is too late. Compound that with the fact that we ourselves live busy lives and have our own selves to take care of, it can be hard to try and look for signs of depression in others, like family members and loved ones.
This is where AI comes in handy and over at MIT, researchers have actually developed a neural network that has the ability to detect signs of depression based on how we talk. This model can be applied to all sorts of conversations, such as interviews, where it can help to predict if the person is depressed without the need for other information.
Now AI that detects depression isn’t new. We have seen how some researchers have used AI to detect depression from Instagram posts. However one of the differences between existing models and MIT’s model is that the former usually requires patients to respond to specific questions, meaning that it limits how and when the model can be applied.
MIT’s model on the other hand uses speech patterns to detect depression, without the need for specific questions and/or answers. One of the potential uses of such a model is that it can be applied to all sorts of conversations, meaning that it could possibly one day exist in an app on your smartphone to monitor conversations and send alerts when it detects something is amiss.
Privacy concerns of such a feature aside, it’s actually not a bad idea, although whether or not we’ll actually see it implemented remains to be seen.