AI belongs on the street alongside cops, researchers urge

‘RoboCop’ was brutal satire about policing, but a team of researchers at a U.S. national lab may see it as the future of law enforcement, according to the digital rights advocacy group Electronic Frontier Foundation.

An autonomous artificial intelligence that can take the wheel during chases, scour social media for fishy activity and generally act as a functional teammate — this is a vision advocated by researchers at the Pacific Northwest National Laboratory, which is managed by the non-profit Battelle Memorial Institute on behalf of the Department of Energy.

A presentation available online describes digital police officers, or so-called D-POs, as “not simply tools we use, but teammates that we work alongside,” non-human officers that “play an active role in the mission.” The deck includes concept art and a scenario in which a D-PO assists a human officer in attempting to stop a robbery and is ultimately told by the human to take control of the operation. Drones, facial recognition and predictive algorithms all enter into the equation.

In another scenario, a “digital teammate,” who in this case is a human, in an augmented reality helmet plays a support role during a border patrol inspection.

At least two of the authors of the report have worked with others at the national lab to publicize the idea. An article on it was carried last November by publisher DomesticPreparedness.com.

Some of the same names are attached to a 2020 article proposal to Fire Chief magazine to push the idea of putting AI out with emergency medical workers, according to the presentation.

The idea presented at the federal lab joins a host of other law enforcement strategies that would employ biometrics and facial recognition.

The authors admitted in emails obtained by the foundation that their vision is still just that, although police in Handan, China, put AI surveillance cops on streets in 2019 to run the faces and plates of drivers.

The ideas have raised concerns among citizens and industry watchdogs about the potential for bias and abuse. In the D-PO fantasy, the virtual partner uses “deep learning image recognition” to make a “high-confidence match” of a suspect. The proposal does not specify whether the system makes use of facial recognition technology — but chances are it is not working with the typical “male, six-foot, red sweatshirt” model.

Whether it is facial recognition or robot dogs, the question hovering over proposed technology for law enforcement is the same one Peter Weller’s character Murphy/RoboCop faced in director Paul Verhoeven’s movie: Can virtual cops serve the public trust, protect the innocent, and uphold the law?

Article Topics

AI  |  biometrics  |  drones  |  facial recognition  |  law enforcement  |  research and development  |  U.S. Government