There’s a new sheriff in Times Square … and it’s an NYPD robot

Millennials heading home from the bar at around midnight last week stopped in the concourse of the Times Square subway station — dollar slices in hand — to take selfies with a 420-pound robot.

The NYPD’s Knightscope K5 Autonomous Security Robot has been making appearances in the station since September, when Mayor Eric Adams declared it would act as a crime deterrent and provide real-time information about where and when to deploy officers to a crime scene.

It’s shaped like a penguin and can’t go up and down stairs. But it has a 360-degree camera on its head and a coin-sized button that connects to a human dispatcher for 311. Video from the robot’s cameras is streamed to the NYPD headquarters in lower Manhattan.

The robot’s cameras add to the many that already exist around the Times Square subway station. “We have more cameras than a Las Vegas casino,” MTA President Richard Davey said at the robot unveiling in September.

The K5 acts as an extra set of eyes and a roving call box, Davey said. Eventually, police will use the robot to respond to 911 calls inside the station and get “a closer look into what’s going on,” said Vladimir Boguslavsky, a detective in the NYPD Transit Bureau specializing in research and development.

The city is leasing the robot from the security and robotics company Knightscope for $9 an hour, which, according to Adams, is more cost-effective than paying human transit cops. Currently it’s escorted on its midnight-to-6 a.m. shift by two police officers, but eventually it’s expected to patrol alone, Adams said.

“This is below minimum wage,” he said when announcing the robot’s two-month pilot program. “No bathroom breaks, no meal breaks. This is a good investment.”

NYPD officials said the robot does not record audio or use facial recognition technology, though Knightscope’s other robots — some of which are used in airports, hospitals and casinos across the United States — do have facial recognition software, per the company’s website.

“We are still in the first phase of testing,” Boguslavsky said. “We intentionally wanted to take our time with the rollout of a technology that’s never been used inside a subway system of this size.”

Boguslavsky said it’s unclear when the robot will begin patrolling solo. He added that the robot is still learning its way around the station and that NYPD officials share feedback and review software performance with Knightscope weekly.

Founded in 2013, Knightscope has deployed its K5 robots to at least one small police department in Huntington, California, according to its website.

They’ve helped a casino operator save over $100,000 “by avoiding a lawsuit over a slip and fall incident using eye-level video evidence,” kept nurses and doctors feeling safe while walking to their cars after dark, and provided high-definition video evidence of a physical assault, Knightscope said on its site.

The robocops have been rolling around LeFrak City, a 20-tower apartment complex in Corona, Queens, for six years and recently were deployed to patrol Lowe’s parking lots and a storage facility in California to deter theft.

But the robots have had some blunders. In 2016, a K5 stationed in a shopping mall in Palo Alto, California, hit a 16-month-old boy in the head and knocked him over. (He was not seriously injured, according to a news report.) One year later, another K5 drowned in a water fountain at an office complex in Washington, D.C.

In New York City, commuters seemed more interested in starting conversations with the robot than relying on it for protection.

“I’m more likely to call 911 if I feel like I’m in trouble than I would run up to the robot and smack it in the face,” said Elias Faris, a 19-year-old Manhattan native, of pressing the robot’s 311 call button.

Another commuter, Maggie Malone, said the robot did not quell her fears of being thrown onto subway tracks. “I’d rather have a person than a machine,” she said.

A Gothamist reporter tested the 311 call button, but the dispatcher on the other line could not hear anything.