An op-ed on the pitfalls of algorithmically-targeted advertising, originally from the Washington Post <https://www.nzherald.co.nz/business/news/article.cfm?c_id=3&objectid=12176094>:

Please, Tech Companies, I implore you: If your algorithms are smart enough to realize that I was pregnant, or that I've given birth, then surely they can be smart enough to realize that my baby died, and advertise to me accordingly - or maybe, just maybe, not at all.

Unfortunately, the old adage “if you’re not paying for the product, then you are the product” applies yet again. Notwithstanding a personal reply from a high-up Facebook exec at the end of the above article, don’t expect these algorithms to be “smart enough” to do anything more than maximize advertising revenue. Anything more ends up counting as “unnecessary expenditure”.

I hate the term "smart" (and AI as well)... Any (real world) model will have false positives. Trying to achieve a higher accuracy/better performance (eg accounting small percentage of cases due to low infant mortality) comes with significant costs, which will eat into your revenue. In other words, improving beyond a certain point is of no interest to the advertiser. Simple cost/benefit analysis for them. Sadly still heart breaking for affected people. Cheers, Peter -- Peter Reutemann Dept. of Computer Science University of Waikato, NZ +64 (7) 858-5174 http://www.cms.waikato.ac.nz/~fracpete/ http://www.data-mining.co.nz/