This is the idea behind the nascent "mosaic" theory of the Fourth Amendment: that by aggregating together many individual observations that each in themselves aren't a "search," you arrive at a result that gives you so much information about a person and their activities that the aggregate program would be a search that required a warrant. Put another way, each individual "tile" of a search program might not be so controversial, like police being able to take your photo in public, but a mosaic formed by aggregating photos of you automatically taken every block and matching them to your drivers' license photo becomes far more problematic, privacy-wise: the whole is more invasive than the sum of its parts.

The theory has made a minor appearance in at least one Supreme Court case: U.S. v. Jones, where police placed a GPS tracker on a car and didn't comply with the restrictions of the warrant they'd obtained to do so. The holding went against the government, but the mosaic theory only featured in a concurrence. The decision was based not around the invasion of privacy inherent in round-the-clock automated location monitoring but instead on the rather mundane and case-specific fact that the police committed a "trespass" to the car by sticking the device magnetically to its underside.

It's a tough case because it's really the plummeting cost that's changing things: round-the-clock warrantless public surveillance is legal and basically uncontroversial if done the traditional way, by sticking police officers in a car and following you around. But the growing ability to get the same information at massive scale, at minimal marginal cost, and even retroactively is... not so uncontroversial. The Court, and courts generally, are still firmly in the camp of analyzing searches individually and not in aggregate, but there are glimmers of change on the horizon.