Whether you think surveillance is necessary or an intrusion, automating it for large populations reveals serious flaws. If a test can produce a false positive, i.e. say something is true when it is not, then the test is fallible. Such imperfections are inherent in a chaotic system, and are therefore unavoidable. That’s why each item flagged by surveillance should be dispassionately reviewed. Unfortunately, even small populations can generate so many false positives that it becomes unrealistic to review them all properly. In the UK, a 1 in a 1000 false positive rate would generate 60,000 alerts. More so for the US. Even a lower false positive rate can generate thousands of alerts against innocents when a system is applied to a population of hundreds of millions. More data is not always the answer.