|
Authored by: stegu on Wednesday, January 09 2013 @ 03:37 PM EST |
> For stuff like this to work you need a database with lots
> of people (everybody would be nice).
We are well on our way to having that, in full public
view. Blogs, tweets and forum posts are voluntarily signed
by traceable individuals, and those sources are very
easily mined for language patterns.
Keep in mind that there is no real need to identify
the author of each individual text with certainty.
Hints and probabilities can be enough to connect the
dots if needed. Besides, what has me worried is not
that I can be identified - I fear more for being
"profiled", lumped together in a focus group and
treated not as an individual, but as a stereotype.
I suppose we could communicate in a deliberately
small, rigid and regular subset of natural language
to remain anonymous, but that would take the fun out
of both writing and reading.
Imagine a paperclip popping up while you are writing:
"I see you are using big words. Big words are more
likely to make it possible to identify you as the
author of this text. Would you like me to suggest
smaller words you can use instead?"
[ Reply to This | Parent | # ]
|
|
|
|
|