Yesterday, it was the news that these deals gave Amazon PR a veto over public statements about the program, and banned cops from disclosing that they were being given free merch and other incentives to recommend Amazon's products.
Now, further reporting from Gizmodo reveals that Amazon's deals also include realtime access to the 911 system, so that Amazon can repackage 911 alerts as "alerts" that are pushed out through the "Neighbors" app. The app is intended for use with Ring surveillance doorbells, but can also stand alone, and allows people to form a kind of curtain-twitchers' alliance with others nearby to combine their (often unfounded) suspicions of (often brown-skinned) passers-by with footage and stills from their home surveillance devices, and then discuss them in a forum that is like Facebook for frightened people.
The problem is that crime is in steady decline, and even the most motivated, toxic pecksniffs can't really manufacture enough content to make Neighbors into a compelling proposition.
The agreements between police and Amazon give Neighbors access to information provided to 911 operators and fed into the computer-aided dispatch (CAD) system, which can include extremely sensitive information ("names, phone numbers, addresses, medical conditions and potentially other types of personally identifiable information, including, in some instances, GPS coordinates").
The internal docs show that Amazon's "Neighbors News team" processes this data and turns it to "alerts" that are pushed to Neighbors users, based on their proximity to the incidents, for crimes from eight categories: "burglary, vehicle break-in and theft, robbery, shots fired, shootings, stabbing, hostage, and arson," as well as "residential, commercial, and structure fires, as well as explosions."
In some territories, 911 calls are matters of public record, but these are in the minority.
The document also includes a list of more than 60 crimes for which Ring will not alert its users. Among “many more,” these include assault, rape, theft, bomb threats, school lockdowns, trespassing, vandalism, domestic disputes, and “dead person.”
“We use an API call, and ingest applicable categories and data points into our content management system, and all incidents will be edited and reviewed before being sent to our users,” the company told police in its marketing material.
It further states that the incidents “must be timely.” Ring will not issue alerts, it says, for crimes that are more than an hour old. “The incident must be useful for the community at large,” the materials say. “If it only affects an individual and does not pose as an active threat to others, we will not post.”
Cops Are Giving Amazon's Ring Your Real-Time 911 Caller Data [Dell Cameron/Gizmodo]