Home improvement products retailer Lowe's has started using security robots manufactured by Knightscope in four stores in Philadelphia.

The robots, K5's – first launched in 2015 – are supposed to help the retailer collect evidence in case of criminal prosecutions, and act like “security guards.”

Even though the K5 can detect persons, has 16 microphones, and Light Detection and Ranging (LIDAR), as well as sonar sensors, along with four 360-degree high definition wide angle cameras, it does not have facial recognition software baked into it.

(But Knightscope offers that feature – “seeing” a person, and knowing who it is – in the K1 Tower and KA Hemisphere models.)

K5's other features include a security button on the back that people can press to summon help, what seems like limited communication letting users send customized messages through it and, as it patrols parking lots, the robot can emit a sound.

By using the robots to identify license plates and mobile devices, Lowe's is able to cross-reference that data with an existing database of previous “offenders.”

According to Knightscope execs, the robot looks for a pretty broad range of “known threats” – from persons with existing criminal trespass records and domestic abusers, to sacked employees.

Lowe's has been deploying K5's via a renting arrangement since February in parking lots to identify license plates and any thermal anomalies (which indicate a person should not be in a place at a certain time of day) that are reported in real time, and pays $6-9 per hour for the robots, which have been mockingly nicknamed, “snitchBOTs.”

The local press reveal that it's a play on words referring to “hitchhiking robot” hitchBOT.+ that met its inglorious end in Philadelphia, where it got “decapitated” – after two years of successfully traveling through Canada and Europe.

The implication is that Philadelphians are not exactly robot-friendly, but up to date, no K5's have been destroyed in that city. There has reportedly been an incident when somebody repeatedly hit one with their car (a Reddit user pinned the blame on the robot, though).

Elsewhere, this model has faced some hiccups in interactions with humans. A child was hit by one in Palo Alto (this time, the maker, Knightscope, blamed the child saying it had “ran backward” into the robot).

Another K5 got “toppled” a few years later, and there's also political drama involved: In San Francisco, SPCA was accused of using one to scare off the homeless (the organization said its K5 was there to prevent vandalism and burglaries.)

But – try to “mess” with it, and Knightscope will, apparently, mess with you.

“We have had people attempt to mess with our robots in a similar way that hitchBOT was and we have actually prosecuted those people to the fullest extent,” is how the company's executive VP and chief client officer, Stacy Stephens, put it.

And even though the K5's deployed by Lowe's in order to “heighten the security” of the retailer's locations are safe for now, it doesn't mean that customers are thrilled to see them.

The machine is 5-foot-tall, weighs 400 pounds, and is shaped like an egg.

CyberLink is producing a robot that has potentially far more sinister capabilities than K5, including sophisticated biometric tracking, but this company is making its products “look and feel” warm (literally) and cuddly – and the marketing goes all the way to naming the thing “Lovot” (a word blend from, “love” and, “robot”), and promising the small, Teletubby-like, not physically threatening machine is “a companion that fills owners' lives with comfort and peace of mind.”

But this “sense of comfort” is there to mask a powerful biometrics-powered surveillance machine, that, according to Groove X (who partnered with CyberLink to produce it) got in demand in elementary schools and daycare centers during the pandemic.

Here are some of the features included in the “cuddly” robot, which, according to its makers, is designed to appeal on people's emotions: 50 sensors and “advanced AI features” including deep-learning real time decision making, and then, “an internal heating system that makes it warm to the touch; can recognize household members (personalized interactions trigger AI facial recognition for different household members), can identify individuals at an angle, with identity processing of 0.2 seconds resulting in 99.7 percent accuracy.”