Data analysis for social good

Recently many researchers, whole universities, even companies started to talk about the new topic of "social good". Examples are numerous: "machine learning for social good", "algorithms for social good", "AI for social good", basically "anything which you could imagine for "social good".

But about "which social good" we are talking about and "why now"?

For the first question, there are many different answers. The field of socio physics existed for quite some time - researchers have been modeling social behaviour for long time for now. [1]. But the problem always was - lack of data, or to be more precise, lack of labelled high quality data. Recently the situation dramatically changed.

Why now? There are also many reasons for that: after famous story with Cambridge analytica many people got to know that the fact that they liked some random cat picture could be used not just against them, it can be used against. Then since many people got scared then there have been many different new laws, e.g. many of you got emails about GDPR from Google, Facebook, etc. where they were stating about the privacy etc.

For scientists however these changes made a big problem. People got scared to donate their data. At the same time various organisations and even movements started to appear around this topic, to name but a few, AInow institute, communities [3] etc.

[1] https://arxiv.org/abs/1207.6178

[2] https://arxiv.org/abs/1210.0137

[3] Data for social good correlaid.org