Machine learned cruelty and border control
The news about immigration control in the US are a daily display of cruelty. Children separated from their families, held in cages, detained in conditions that would be considered inhumane for animals, let alone vulnerable children.
Reporters given access to a facility at the southern border, where the US holds arrested families, find traumatised children inside cages without toys or books and with large foil sheets for blankets. Photo: AP https://t.co/UDnZlECbQy pic.twitter.com/JA3OviNYp8
— ABC News (@abcnews) June 18, 2018
US Border Patrol: Hundreds of children kept in cages at facility in Texas
and this thread for the European similarities
so when I see the photos and the purposeful cruelty now taking place in the US, I do not see them as exclusively Trump's doing. He has only dialed up what Obama had already been doing and has further expanded what has been going on in the EU for decades.
— Flavia Dzodan (@redlightvoices) June 19, 2018
Elsewhere I wrote:
[…] today, capitalism depends on a strategy of habitual cruelty for the purpose of numbing us to its effects. The current paradigm of exploitation requires a wide spectrum of vulnerability of human life. In this context, harassment is the strategy. It’s not merely the tool to create and promote a state of helplessness and vulnerability but it is at the very root of a widespread system of violence on which the foundations of exploitation can be laid down.
as I have said before elsewhere, harassment and violence are not bugs [in the system]. They are features in a system that demands docility and compliance to sustain policies of austerity and marginalization. As an overlay to the biopolitics of control, the necropolitics of capitalist evaluation of life’s worth and livability which necessitates this constant use of violence to exert compliance.
And here I need to point out to the way a new pedagogy is emerging: machine learned cruelty which I’d tentatively call “a pedagogy of robots” in the sense of teaching algorithms how to detect migrants so that ICE can more effectively and swiftly detain them while in transit.
From a feature at Wired last week: INSIDE PALMER LUCKEY’S BID TO BUILD A BORDER WALL – How the Oculus founder, along with ex-Palantir executives, plans to reinvent national security, starting with Trump’s agenda.
Within a couple of months, Anduril had a prototype. Schimpf and his colleagues took it to a test range in Apple Valley, a two-hour drive from their Orange County office. “We lived out of the trailer there,” Schimpf says. Using open source machine-learning training data, they taught the software how to tell humans from animals or tumbleweeds, and unearthed some glitches. In a certain light, for example, the system can mistake the rear end of a horse for a person.
This “pedagogy of robots” is deployed on military applications, border control technologies and the instruments of bio and necropolitical control. As an extension to the pedagogy of cruelty, the technological applications that automatize inhumanity. Paradoxically, while creating more human machines, algorithms are transferred the most inhumane aspects of being human. “Machine learning” then, as an extension of practices of subjugation and brutality.
and the efficiency of the system is measured in number of arrests:
In one Lattice demo, author Stephen Levy put on a Samsung Gear VR headset that showed the wearer a direct video feed of the border. If anything — a human, vehicle, or animal — tried to pass, the system gave users a bright green alert identifying it along with a probability certainty. The system is currently being tested on a Texas rancher’s private land. Over a 10-week span, Anduril’s security towers apparently helped border agents catch 55 people and seize 982 pounds of marijuana, although 39 of those arrests weren’t connected to drugs.
Microsoft further providing infrastructure to sustain the necropolitics of migration control through machine learning:
This ATO is a critical next step in enabling ICE to deliver such services as cloud-based identity and access, serving both employees and citizens from applications hosted in the cloud. This can help employees make more informed decisions faster, with Azure Government enabling them to process data on edge devices or utilize deep learning capabilities to accelerate facial recognition and identification.
The algorithm as an extension of State policies and, expanding the boundaries of what constitutes the State, it is private corporations and landowners who allow the infrastructures on their property and who are in charge of implementing migration control. Cruelty, it seems, is everybody’s profitable business.
