I guess they are not the first. There is a company from Israel that deployed 10 years ago algoritms to score if people was lying, or angry when people called t

I guess they are not the first. There is a company from Israel that deployed 10 years ago algoritms to score if people was lying, or angry when people called the police. And it makes sense. Anything that helps you triage when you have limited resources it's a good idea. And any tool that suggest deterrent actions to keep your resources is also a good idea (ie police near a metro station next to a stadium hours before a sports event so pickpocketers don't come nearby). For me the weakest point is not the technical part (mechanisms), is how you use them (policies). I mean, you need some policy to trigger reactions to signals. And since policies are made by human people, that has more bias than any algorithm you can ever produce, plus and policy changes will surely affect how the "system" works and how it needs retraining. So, unless this is used for very simple stuff with simple policies it won't work.