There's a common pattern of suggesting that if an employer (or state, or university, or...) can access data, there's no problem when they do. I think it's a pretty serious error.

Broadly, automated surveillance creates illusory offenses. It raises issues based on some imperfect set of rules, and does so independent of any report of harm. (Add to that the likelihood of unethical flagging - which happened right in the article!)

And that creates an obligation to act for employers. Ignoring alerts, no matter how silly, is the sort of thing that might get cited as negligence when a real problem does come down the pike. At best, the result is time-consuming investigations of fake issues. At worst, corporate makes a decision that summary firings are an easier bet than proper investigations.

That's not just a hypothetical fear - we already hear about things like Amazon warehouse employees being laid off on vague suspicion of theft, because investigations are harder than replacing staff. Expanding always-active surveillance will predictably hurt innocents who already lack job security.