What they’re describing is the kind of thing where the “last-gen” iteration/definition of AI, as in pretrained neural networks, are very applicable - taking in various vitals as inputs and outputting a value for if it should be alarming. For simple things you don’t need any of that, but if you want to be detecting more subtle signs to give an early warning, it can be really difficult to manually write logic for that, while machine learning can potentially catch cases you didn’t even think of.
What they’re describing is the kind of thing where the “last-gen” iteration/definition of AI, as in pretrained neural networks, are very applicable - taking in various vitals as inputs and outputting a value for if it should be alarming. For simple things you don’t need any of that, but if you want to be detecting more subtle signs to give an early warning, it can be really difficult to manually write logic for that, while machine learning can potentially catch cases you didn’t even think of.