From the article: [...] (They advocate for the differential privacy approach.)
From the article:
The first step researchers took in breaking the anonymity of this data set feels surprisingly innocuous: they found their own cards in the data set. This is simple: you can look up several times and places where you took a bus. [...]
While identifying your own card scarcely feels like a privacy violation, it enables a much more damaging second step: identifying the card of someone you know. For example, knowing which card in the data set is yours, you can easily identify the card of a coworker. [...] Having identified the card of your coworker, you can find out any other trip they have taken: weekend trips, doctor visits, or other excursions that they probably expect to be private.
The researchers also showed that such privacy breaks can be extended further by linking the data set with publicly available information. [...]
[...]
Let us draw several lessons from this attack:
Anything is potentially personally identifying information [...]
Who does the identification matters. [...]
De-identification is hard in a connected world. [...]
(They advocate for the differential privacy approach.)
From the article:
[...]
(They advocate for the differential privacy approach.)