Re: re-identifying anonymized data
If it can be re-identified, then it hasn't been properly anonymized in the first place.
That is not necessarily the case. Let's say that company A provide anonymized purchase data to a company Z who carry out data analysis.
And then company B provide anonymized health data to company Z,
And then company C provide anonymized travel data to company Z.
Company Z may, through intersections between data from A B and C, be able to identify individuals, where that would be impossible from any single one of the data sets.
That doesn't mean that the individual data sets are not sufficiently anonymized, just that accumulation of many data points from different sources can allow correlations which lead to the identity of the subject.