“The big secret of the industry is that none of those companies has any idea how to make [domestic] robots smart enough to be generally useful. … There are a bunch of breakthroughs that need to arrive in A.I. before that’s possible. So the future of a lot of those companies essentially depends on whether we’re going to make significant progress on those world model/planning-type architectures.” —Yann LeCun, still ostensibly Meta’s chief A.I. scientist, speaking at an M.I.T. symposium about the challenging road ahead for the humanoid robotics industry.

And now for the main event…

Earlier this month, the Miami-Dade Sheriff’s Office unveiled the newest member of its police force: a “Police Unmanned Ground Vehicle,” or PUG, an A.I.-enhanced police cruiser outfitted with self-driving technology developed by Perrone Robotics. The PUG can access real-time crime data and A.I. analytics; is loaded with 360-degree cameras, thermal imaging, and automated license plate readers; and has a drone lashed to the roof. A press release referred to it as a “force multiplier”—a phrase that crops up over and over among the hundreds of initiatives to infuse A.I. into police departments around the world.

Yet it’s still not clear what the PUG will actually do, as numerous industry sources have pointed out. For now, that’s by design: The vehicle is the centerpiece of a 12-month pilot program that will, according to a statement from Miami-Dade Sheriff Rosie Cordero-Stutz, allow the department to “explore how new technology can keep people safe while making the best use of our resources.” Whether the department actually deploys the vehicle after the trial is “going to be a decision made with the sheriff,” said Edward Prokop, a strategic advisor with Policing Lab, the nonprofit that designed and donated the car. Prokop added that the idea behind the car is to enable more “visible deterrence,” and that, “obviously, the comfort level over the next 12 months, with community acceptance and effectiveness of this thing, is going to drive Phase 2.” (The sheriff couldn’t be reached for comment.)

Presented as the first autonomous police cruiser, the PUG is merely the latest example of law enforcement’s appetite for algorithms and analytics to enhance their capabilities. The consequences are often messy: Both the Chicago and Los Angeles police departments scrapped predictive policing initiatives amid public outcry—the L.A.P.D. program had been developed in partnership with Palantir—though they remain committed to “data-driven” crime prevention. And interest in tech-enhanced policing hasn’t diminished despite all sorts of legal scrutiny over its potential biases and inaccuracies. Axon, a company that supplies police departments with tasers and related equipment, recently started selling a suite of A.I. technologies (also advertised, of course, as force multipliers) intended to boost efficiency, help officers spend less time on paperwork, etcetera. The tech is catching on, and moving, as it always does, faster than regulatory groups can keep up.

To envision the PUG as merely a computer on wheels doesn’t really do it justice. Anthony Grubisic, a partner and treasurer at Policing Lab, told me the car features a full booking station, replete with a facial recognition camera, palm and fingerprint scanners, and an ID reader that can spot a fake. He also said they’re looking into adding a computer panel, which would allow people to interact with some sort of A.I.-based application. In other words, it’s basically a fully automated paddy wagon—although for the length of the trial, the vehicle won’t go anywhere without a deputy in the driver’s seat, despite its Level 3 self-driving capabilities. The PUG will travel on preplanned routes to locations that have been identified by the sheriff’s office as needing “more visible deterrence,” while still receiving specific missions from the office.

Nevertheless, and predictably, the robo-cruiser’s comprehensive suite of information-gathering capabilities has raised alarms among critics. Ivey Dyson, an attorney at the Brennan Center, which has filed multiple lawsuits over data-driven policing, argued that the PUG “should not just be viewed as surveillance technology. … [It] should be viewed as a mass data collection technology.” Dyson explained that this next generation of policing tools often “raises privacy and civil liberties concerns,” but that many police departments lack the capability to properly manage them. Thermal imaging, for example, might mistake a security guard for someone with unauthorized access to a given area, which “could lead to an excessive police response to generally harmless activity,” she said. Predictive policing programs have also run into trouble in the past thanks to algorithmic bias that disproportionately affects minority communities.

But because police departments are typically “overwhelmed and under-resourced,” Dyson said, they often defer to the equipment vendor rather than tackle those problems themselves. And it’s hard to look a gift horse in the mouth. Nobody would give me a price tag for the PUG—Prokop said it’s “hard to put a dollar amount” on the development costs—but it’s hard to imagine all those bells and whistles come cheap. Thankfully for Miami-Dade, Policing Lab picked up the bill.

Perhaps it’s the nature of any pilot phase, but at this point, the PUG seems to raise more questions than answers. Catherine Crump, an attorney and law professor at UC Berkeley, told me that these questions are vitally important for the public to consider before the PUG is officially launched. “The public should ask what exactly the car is doing, why it’s better at these tasks than an officer would be, and what the privacy and civil liberties impacts will be,” she told me. “What is going to happen to the data this vehicle gathers? How will it be used? How long will it be stored? And before it’s deleted—assuming it isn’t kept forever—who will it be shared with?”

For his part, Grubisic told me that the data collected so far has been stored on a hard drive within the car. “That comes with us. We don’t leave it in the vehicle, and the data is erased right then and there,” he said. “But we’re not collecting any public information in any way at this point.” Grubisic added that Policing Lab is working with the sheriff’s office to develop a cloud-based “backbone” to support information management. Meanwhile, Prokop acknowledged the privacy concerns, but said that the PUG doesn’t possess any surveillance or technological capability that the department doesn’t already have.

Crump conceded that new policing technologies are not inherently problematic, and that “there are times when taking the human out of the equation can promote safety.” But until they’re fully understood and vetted, these technologies can introduce a host of complications. Jon Mills, a privacy expert and University of Florida law professor, pointed to clashes over Fourth Amendment protections that have accompanied police drone initiatives, particularly regarding the limits of warrantless surveillance. Ultimately, he said, it’s up to law enforcement to “sensibly limit” the use of these technologies and apply discretion—a notion that might not sit well with communities that have been affected by police overreach.

Of course, given the undisguised excitement many departments have toward new technologies, tools like the PUG will likely continue to be rolled out. “A.I. is a shiny new toy,” Dyson told me, before adding, on a perhaps unintentionally ominous note: “Police departments might not have the capability to really understand how it works.”

That’s all for today. I’ll see you on Thursday.

Ian

Need help? Review our FAQ page or contact us for assistance. For brand partnerships, email ads@puck.news.

You received this email because you signed up to receive emails from Puck, or as part of your Puck account associated with {{customer.email}}. To stop receiving this newsletter and/or manage all your email preferences, click here.

Puck is published by Heat Media LLC. 107 Greenwich St, New York, NY 10006