[Sigia-l] Will she be fired?

[Listera]

Laurie Gray:

> Do you see a backlash occurring?

Hmm. Only if we have the equivalent of three or four 9/11s in the data
privacy area. 

The system I'm re-designing at the moment (and others in the field) can
paint an amazingly accurate and detailed picture of your financial
background given just a few rudimentary facts about you. It can also predict
your financial behavior into the future with uncanny precision. It's not
infallible obviously, but it's not voodoo either.

Its predictive analytics, however, couldn't function without historical
data. Not personal financial transactional data per se but aggregated
statistical data about a representative population. Yes, you can gain
further insight by mapping this to an individual's personal history but
that's less essential than it might first appear. The risk without personal
info is manageable, but without statistically significant aggregate data
it's not.

There are some truly delicious system architecture challenges I have to deal
with but the pertinent issue here is the significance and the integrity of
personal data.

ID theft and fraud are a huge issue for financial services providers. Right
now the general public is paying for what's regarded as "acceptable"
default/risk rates. It might surprise you a little bit in that those in the
industry who can see a little further into the future are resigned to the
fact that given our present course fraud is not preventable. They are now
working on systems :-) that will assume/accept (high rate of) fraud and
route around it.  

As in medicine or education, predictive analytics can really help
individuals make better financial decisions for themselves, even as they
help companies create truly better-fitting products. But that can only
happen when aggregated/historical data is available. (BTW, this also happens
to be the basis of a variety of social networks/apps.) It would thus be a
shame if the pendulum swings (totally) the other way and eliminates the
possibility of gathering such data.

----
Ziya

Dare to do less.