Can AI Really Predict Crime Before It Happens?
The short version: SciVault's dive into predictive policing AI is a reminder that ALPR networks like Flock feed the same data pipelines these systems depend on.
Our Take
SciVault's explainer on predictive policing AI raises the right question: can algorithms really forecast crime, or are they just dressing up old biases in new math? We'd add a piece the video only gestures at — these predictive systems don't run on magic, they run on data. And a huge chunk of that data increasingly comes from automated license plate readers like Flock Safety's, quietly logging every car that passes a camera, building location histories on millions of people who've never been charged with anything.
That's the part that should worry you more than sci-fi crime prediction itself. When police departments feed ALPR logs into 'predictive' models, they're not forecasting the future — they're reinforcing whatever patterns already exist in over-policed neighborhoods, then calling the output objective. The tech sounds neutral. The inputs rarely are.
Credit to SciVault for laying out the AI side clearly. On our end, we track where the cameras feeding these systems actually sit — check the map to see what's watching your own neighborhood, and hit take action if you want to push back on it.
This is DeFlock The USA’s original commentary. The video above is the work of SciVault, published on YouTube — full credit to the creator.