Earlier quoted context omitted.
In my view, the state has to ask a court to approve the purpose of each request, based on a reasonable hypothesis about what might be found. How this seems different is the lack of any limitation on the kind of content that can be turned over. And in my view, that was a mistake on the judge's part for approving such a sweeping request. For instance if I were accused of committing a crime against X, then I suppose tha…
How do you propose that works? Someone at Google searches for the information? An algorithm or an actual Google employee? If it is a human being then what about your privacy then? What if the Jussie and his manager were using codewords to disguise intent? A search algorithm would not be able to recognise that but a human being can easily infer meaning and intent from the context. What then? There are no easy answers.
This distinction wasn't a major concern when the former was impractical.
Because of digital records and search agents, suddenly either is practical.
The intent of privacy law has always been to raise the cost of targeted prosecution. E.g. "We want to charge this person with a crime, let's find a crime."
The state absolutely has a duty to prosecute and hold individuals accountable. And in extraordinary circumstances, may even need to target prosecution.
What is unacceptable in a free society is that the state should have the ability to target anyone for prosecution with no effective bounds in the number of simultaneous times it does so (aka everyone).
The difference between Orwell and a safe democracy is scale.