This *looks* like a pro-social use of AI, but it's not https://www.thetimes.com/uk/politics/article/ai-data-children-predict-criminals-fwclzh323
1. Sociology has a pretty good idea of what factors contribute to crime, such as poverty. This "solution" is a surveillance-based intervention into individual rather than structural harms.
2. It fails to notice that predictive risk systems themselves victimize the vulnerable; they amplify bias and create feedback loops.
3. This will necessarily treat children as pre-criminals.
4
