Siamo usciti dalla fase sperimentale: l'applicazione delle tecnologie della sorveglianza per l'arresto e la deportazione degli immigrati prefigura il nuovo scenario per tutti. <https://www.nytimes.com/2019/10/02/magazine/ice-surveillance-deportation.htm...> [...] Tracking immigrants in this country is an increasingly trivial exercise because it’s an increasingly trivial exercise to track any of us. The same phenomenon helped make the reporting for this article possible: Over the course of more than a year, I tried to reverse-engineer individual ICE officers’ use of America’s vast post-Sept. 11 domestic-surveillance apparatus, retracing their hunt for targets down to the very searches they entered into their computers. Based on hundreds of government-procurement records, dozens of interviews and thousands of pages of documents I and others obtained via public-records requests, the account that follows reveals new evidence of surveillance of detainees’ voice and video calls at ICE facilities and extensive proof that the agency relies on state D.M.V. databases and information products like CLEAR, from Thomson Reuters, to target immigrants. It shows how the Real ID Act of 2005, along with funding from the Department of Homeland Security (D.H.S.) and help from umbrella groups like the American Association of Motor Vehicle Administrators (A.A.M.V.A.), has laid the groundwork for real-time, real-world monitoring of American residents. It’s still a partial view. But it’s enough to show the scale of the machine. ICE’s transformation began under President Barack Obama. Criticized by activists as the “deporter in chief,” Obama indeed removed more than three million people from the country, including, in 2013, what remains the annual record — 432,448 people. But part of the spike he oversaw was an accounting trick: Near the Southern border, migrants caught trying to cross were increasingly subject to formal removal proceedings (and thus officially counted in statistics), a change from previous administrations. Away from the border, Obama tried to shift ICE’s priorities. He curtailed the agency’s longstanding practice of workplace raids, and active pursuit was now largely limited, especially during his second term, to convicted criminals, suspected gang members or terrorists and people apprehended at the border. The administration also discouraged so-called collateral arrests, in which ICE came looking for one person and instead nabbed a neighbor or a co-worker. The lowest priority were long-term residents with American children or spouses. “Felons, not families,” Obama explained in a 2014 speech. “Criminals, not children. Gang members, not a mom who is working hard to provide for her kids.” ICE now needed systems that could more intelligently parse priority targets from the general population of millions of undocumented immigrants. Under Obama, the agency signed several key data and analytics contracts that have since drawn controversy, including with Thomson Reuters, the Silicon Valley firm Palantir and the Beltway start-up Giant Oak, and it relied ever more on local-federal data-sharing initiatives that flourished after Sept. 11. Information that was once siloed or tedious to access — driver’s-license photos, phone records, jail bookings — was increasingly at deportation officers’ fingertips. One of Trump’s first acts as president was to throw out his predecessor’s priority list. He issued Executive Order 13768, “Enhancing Public Safety in the Interior of the United States,” on his sixth day in office. “We cannot faithfully execute the immigration laws of the United States,” the order reads, “if we exempt classes or categories of removable aliens from potential enforcement.” Data once used to focus on prioritized groups could now, in theory, be used against anyone, and individual officers now had a freer hand to pursue their own priorities. The results were predictable: Both the percentage and sheer number of arrests by ICE of people without criminal convictions shot upward. There were 58,010 such arrests during the administration’s first 14 months, according to an analysis by NBC News — nearly three times as many as during the preceding 14 months. Statistics compiled by Syracuse University’s Transactional Records Access Clearinghouse, meanwhile, show a significant rise in immigration cases involving long-term residents. In an interview before her retirement this year, Elizabeth Godfrey, ICE’s acting field office director for the Pacific Northwest region, pointed to another reason the profile of its arrestees had changed: A series of pre-Trump court decisions, notably one made by a federal judge in Oregon in 2014, led a growing number of jails across the country to stop honoring ICE requests that they notify the agency of pending releases or hold immigrants beyond their scheduled release dates. This, together with sanctuary policies in a growing number of jurisdictions, cut off deportation officers’ easiest source of arrests, driving them to new hunting grounds. Officers working out of Godfrey’s Portland office — which was shut down for more than a week by protests last year — still carefully prioritized their targets, she said. They studied them in advance. Before attempting an arrest, they completed a form, the Field Operations Worksheet, and had it approved by two layers of supervisors. When given a choice, they would always pursue someone with a serious criminal history rather than someone without. But that wasn’t always the choice. Targeting decisions also depend on who’s findable. And today who’s findable depends on who has left a trail for officers to follow: people with driver’s licenses, people with car insurance, people with utility bills, people who pay U.S. taxes, people with social-media accounts, people with stable homes, people with American-born children, people who live American lives, people who once thought — if only for a moment — that they weren’t on the list. “We still have a job to do,” Godfrey told me. “We still have a mandate. So where are we going to locate these people? At a place of employment or at their home or anywhere else we can find them.” Image Shellfish beds near the Port of Peninsula, Nahcotta, Wash. Over time, ICE officers began arresting many of the industry’s workers. Shellfish beds near the Port of Peninsula, Nahcotta, Wash. Over time, ICE officers began arresting many of the industry’s workers.CreditMatt Black/Magnum, for The New York Times 3. ‘They knew who I was.’ The first person outside Pacific County’s Mexican community to notice the ICE crackdown was Erin Glenn, a soft-spoken high school teacher whose family has lived and farmed in the area for five generations. In 2015, after years of little noticeable ICE activity, a man she’d known for decades was suddenly picked up and deported. At first Glenn, who studied Spanish in college and later taught English to recent immigrants, let doubts creep in. “It’s the feds,” she says. “I thought they must know something about him I didn’t.” But the next year, as Trump campaigned and soon won the presidency on promises to clamp down on immigration, ICE arrested at least eight more of Glenn’s neighbors and friends, and she saw a trend. She began to privately tally the names of the people who had disappeared. The arrests seemed to accelerate, and the list became more detailed: how people were taken, where they were taken, when they were taken and whom they left behind. [...] If the recurring fear in Pacific County was that some unknown person — the sheriff, a local informant — was secretly behind ICE’s arrests, the recurring fear among rights groups in the other Washington, in the District of Columbia, was quite the opposite: that immigration decisions would soon be so automated as to effectively cut humans out of the loop. Exhibit A as the new administration took control of the capital was the Extreme Vetting Initiative. The initiative was ICE’s attempt to bring an ill-defined Trump campaign promise to life using artificial intelligence. The agency introduced it in a July 2017 “industry day” in a Marriott conference room near Ronald Reagan Washington National Airport, inviting a flock of technology firms and Beltway contractors. The event’s host was the Counterterrorism and Criminal Exploitation Unit (C.T.C.E.U.), an office within H.S.I. that scrutinizes visa applicants and visa holders before and after they arrive in the United States. The guests, according to a sign-in sheet obtained by The Intercept, were a parade of billion-dollar companies — Thomson Reuters, LexisNexis, IBM, Booz Allen Hamilton, Deloitte, PricewaterhouseCoopers — along with various smaller newcomers, including the social-media monitoring start-ups Giant Oak and Babel Street. To build the system it wanted, ICE expected to award one or some of them a contract worth $100 million or more. Interest was so high that the Marriott event, originally planned for one day, had to stretch to two. The goal of the Extreme Vetting Initiative was “determinations via automation,” ICE explained to the attendees. It sought partners whose algorithms could scan social media and other publicly available information and assess whether an immigrant was likely to become a “positively contributing member of society” — or whether he or she intended “to commit criminal or terrorist attacks.” Written text would be analyzed for sentiment, Facebook posts for aberrant thoughts, emojis for signs of criminality. The system would replace C.T.C.E.U.’s human analysts and “generate a minimum of 10,000 investigative leads annually.” In early 2018, ICE seemed to retreat from the Extreme Vetting Initiative, which has since been renamed the Visa Lifecycle Vetting Initiative. The contract, for $102 million, went to CSRA, a Beltway giant already doing work for ICE that has since been purchased by General Dynamics. The company employed more than 100 human analysts — ensuring that people would still be firmly in charge of vetting other people. But the retreat, procurement documents show, was partial at best: H.S.I. had already been testing automated social-media profiling under Obama, and many of the technology companies that showed up at ICE’s Marriott event were already working for it. In 2016, ICE paid a broadband provider $50,400 “to strengthen C.T.C.E.U.’s open-source social-media exploitation capabilities as well as future growth by adding fiber-optic internet.” For the social-media capabilities themselves, it paid $1.4 million that year — and has since paid millions more — to Giant Oak, a rapidly expanding, 30-person analytics firm in Virginia that was founded by the behavioral economist and former C.B.P. chief of staff Gary Shiffman. One of Giant Oak’s early contracts was for a DARPA “quantitative counterinsurgency” project in Afghanistan. Two years ago, Forbes reported on the company’s work with ICE and detailed its flagship product, GOST, short for Giant Oak Search Technology, which allows users to search the dark web and social media to create “a dossier on each individual with everything you need to know, such as web page images and keywords already highlighted.” When I met Shiffman in a pub in Georgetown last year, he confirmed that the company worked for H.S.I., vetting visa holders and visa applicants, but not for E.R.O. Its technology wasn’t in the hands of domestic deportation officers like Miller and Dietz. He made clear that he didn’t want to live in a country that indiscriminately rounded up foreigners. But if public safety was the goal, the answer wasn’t to stop scanning people’s data. It was to do it the right way. “Now this is my theory,” he told me, “and you tell me if it’s crazy or wrong. But I think that the better we have entity resolution” — that is, the better we can compile and measure people’s data — “the less of a surveillance state we’ll have.” Once the machine is perfect, as long as you’re following the rules, you won’t even have to know it’s there. [...]