There's also the fact that the courts are driven by narratives; the prosecution just needs a credible story to tell, and doesn't have much of an incentive to find out the truth - it's up to the defense to tell the counter-narrative.
Over time, as more tools become available to find people with precise overlaps with a crime, the stronger the narrative is for innocent people (as well as the guilty). IMO a more inquisitorial approach which tests the counterfactual of each evidentiary point rather than relying solely on defense ought to hold up better in the future.
I doubt better surveillance tools will lead to more accurate prosecutions unless individuals also have direct access to the tools.
Police and prosecutors in the US are notorious for leaving out evidence that exonerates the accused, even though they are supposedly required to provide the defense with all evidence relevant to the case.
Also, well-connected victims’ cases are investigated much more carefully than normal victims.
For instance, some SFO luggage handlers were stealing stuff from bags. People repeatedly reported them to the police, but nothing happened until they stole from a retired cop. An investigation was launched and they were arrested the next week.
Similarly (based on personal experience with multiple incidents in multiple cities with multiple victims), they normally won’t bother to pull location database records for successful burglaries, even if they have exact timestamps of the incident (up to and including video), even in remote areas in the middle of the night involving literal truckloads of loot (where the location records would be a slam dunk, and the evidence would take a while to dispose of).
Over time, as more tools become available to find people with precise overlaps with a crime, the stronger the narrative is for innocent people (as well as the guilty). IMO a more inquisitorial approach which tests the counterfactual of each evidentiary point rather than relying solely on defense ought to hold up better in the future.