Earlier quoted context omitted.
Detecting plain SSN numbers wouldn't be difficult with a combination of regex, machine learning and human verification. Even if hackers could just encode the SSN numbers, it would at least mitigate the spreading of PII. Edit : I don't care about the downvotes, I care about privacy. Enough of the argument "but wait, can't you imagine the cost?", well if you can't afford to protect people's privacy, don't do business a…
You missed the whole point of the comment you replied to. You can detect the format of the SSN, but you can't reliably tell that it actually is a SSN, and not some other type of identifier.
I think you can. Now when we have neural nets capable of distinguishing 100s of different dog breeds, we still have to trip on the most basic and structured type of entity extraction? No. A simple regex, combined with heuristics, and a linear model on top can reliably detect SSN's.
Just that there will be a few false positives (no matter if you automate this, or do this manually) does not mean it is a Herculean technical challenge to do this.