# In the future everyone will be discriminated against for 15 minutes A few days ago, my boss noticed that Amazon was offering him a much higher price for the same product than us underlings. Yesterday a friend observed that it's possible to reliably estimate someone's political affiliation from publicly available data. If you had Facebook or Google or Amazon's cooperation (and for the record Facebook shares no information on specific individuals, at least that I've seen), it would not be difficult for a store to offer a higher price to republicans, or environmentalists, or people called Dave. As I understand it, current law makes discrimination against specific kinds of groups illegal. But shopkeepers and the like enjoy the right to treat specific individuals differently, for any reason other than membership of these groups; if you want to refuse service to Bob because he looks scruffy, or slept with your girlfriend, that's your right. US employment law as I understand it is stricter: if some factor is correlated with membership of a protected category, then you may only use this factor in hiring decisions if it is demonstrably related to the specific job you're hiring for. (Given that, in the age of big data, correlations can be found between approximately any pair of factors, it seems like this makes it illegal to base hiring decisions on anything other than job-related factors? Which is not unreasonable, but seems like a huge broadening of the intended scope of the law in question). This was probably a pragmatic option at the time. A human shopkeeper can only judge by what they can see; the only categories they can treat differently are those that are clearly visible. It's easy to discriminate by ethnicity or sex or age or disability, and you can make a decent guess at sexuality or transsexuality. Interestingly we don't consider social class to be a protected category - perhaps because it's difficult to define objectively, perhaps because we feel there's something fair in discriminating on grounds of wealth (which is at least closely correlated with class) in a way that there isn't for other categories. (That said, subculture membership is also easy to discriminate by. One can certainly imagine a shopkeeper refusing to serve Goths, or those in hoodies; the former would be socially unacceptable but the latter less so, perhaps because there is the fig leaf that this is about the ease of concealing stolen goods, or perhaps because that particular group is seen as an acceptable target. So maybe these problems already existed to some extent, and we trusted shopkeepers to do the right thing, and communities to hold them to account.) But it's about to get a whole lot worse (or better). The tide of progress means that soon, inevitably, anyone you meet will be aware of all the information publicly available about you - which is quite a lot. Those who want to curate their encounters will know who to talk to and who to avoid. And people who run businesses - which is likely to be more of us in the future - will be able to make finer distinctions than ever before. Refusing to serve someone with a poor credit rating is clearly just good business sense. Refusing because they like different music from you is far more suspect (particularly given the odds of a racial correlation). It'll be interesting to see where the line gets drawn. [Home](/) <div id="disqus_thread"></div> comments powered by Disqus