By Josh Hawkins
New York City is piloting a new AI camera system in its subway platforms that aims to spot danger before it happens.
The technology, still in its early testing phase, doesn’t use facial recognition. Instead, it watches for erratic movements, signs of distress, or agitation — behavior that could hint at something about to go wrong.
The Metropolitan Transportation Authority (MTA) is leading this experiment.
Their goal? What they call “predictive prevention.” According to MTA Chief Security Officer Michael Kemper, the system helps trigger alerts so police or security can respond before a situation escalates.
“It’s not about watching people,” Kemper said in a recent safety meeting. “It’s about watching for warning signs.” That means if someone starts behaving erratically — pacing, shouting, or appearing panicked — the system might alert human responders immediately.
The MTA has already dipped its toes in AI. Back in 2023, it used AI tools to catch fare evaders by tracking when and where people jumped turnstiles.
This new step takes the technology further — using it for real-time safety monitoring on crowded platforms.
Officials stress that the system does not use facial recognition.
That’s a clear move to ease privacy concerns, which always come up when surveillance and AI are involved. The AI isn’t tracking identities, just behavior.
Still, questions remain. The agency hasn’t disclosed which tech companies are behind the pilot or how accurate the behavior predictions are.
In a chaotic subway setting, false positives could lead to unnecessary interventions — or worse, profiling.
Critics worry the effort may cross ethical lines. Is this a tool for safety, or a slippery slope toward over-surveillance?
For now, the AI camera project is in the pilot phase. But it’s clear New York is leaning into smart surveillance.
Whether this system becomes a model for other cities — or a controversial case study in AI overreach — remains to be seen.
