Hacker News Reader

Comment by Inconel | Hacker News Reader

I'm fairly certain I read that this system will also allow citizens to assign trustworthiness scores to friends and acquaintances.

View on HN

Bentham would have never imagined his Panopticon [1] would be implemented in this manner, across the entire society, in digital form.

[1] https://en.wikipedia.org/wiki/Panopticon

> Credit scoring companies in the US and China are already experimenting with data from social media profiles. Counting how many times someone says “wasted” in their Facebook status can help to predict whether they will repay their debts on time, according to a trial by US credit analyst FICO.

Source: https://www.ft.com/content/e8ccd7b8-6459-11e6-a08a-c7ac04ef0... [FT might be paywalled, but I was able to reach it via Google for free]

That's pretty unsettling. I wonder if I'm going to find myself in a similar situation regarding social networking as I found myself in regarding credit cards a number of years ago. For the longest time I didn't have any credit cards since the idea of debt really scared me but I realized I was only hurting myself by not building a credit history. I signed up for a number of cards over the years and put small recurring charges on them that I always pay off in full. Similarly, I don't really participate in much social media, HN is the exception, and I have no FB account, but I wonder if at some point my lack of public facing social media will negatively impact other areas of my life like borrowing, employment, or housing.

Maybe someone enterprising can start a service that automatically posts a bunch of innocuous stuff to a FB profile so I don't have to bother wasting my time with it.

The sort of profiling the parent comment was talking about doesn't work the other way around - it'd be idiotic to punish someone's credit score for not having a social media account or having it set to private/friends-only/un-crawlable. The lack of a social media account might even be a signal of better creditworthiness

I see bots being extraordinarily effective against this type of automated scoring as the sophistication of the scoring bot seems like it will always lag potential creator bots.

Given that scoring bots in mass use will necessarily decrease the dimensionality of the data into fewer, more tractable features, and thus creating bots need only to target the dimensionality-reduction step, rather than reproduce "human behavior."

Which seems rather simple given the variance in natural language and normal human social interaction.

Reply on news.ycombinator.com