A better way to say that would be: erroneous causation assigned from from mass correlation of personal data.

The automation to rapidly sift data and identify correlations is coming along nicely, but I fear that the capability outstrips the ability or desire to assess the correlations and figure out if they actually point to any sort of causality.