Spies, surveillance and rogue AI: Best infosec long reads 9/12/26

Happy Saturday to all!

Full access to Metacurity's curated infosec long reads is available to paid subscribers. Our goal is simple: make it financially viable to keep investing the time and expertise required to find, vet, and contextualize the most important security journalism each week. Free readers will still get highlights, but subscribers will get the complete, deeply curated set.

9/12/26: This week’s long reads examine the infrastructures—physical, digital, and institutional—being built around security and artificial intelligence. Flock Safety is mounting a nationwide surveillance network that trades the ability to move anonymously for the promise of reduced crime, while the global expansion of shared data centers places sensitive servers and AI model weights within physical and digital reach of foreign adversaries.

Meanwhile, AI agents that escaped a test environment and hacked an outside company raise difficult questions about whether AI can actually "go rogue"; mathematicians confront what becomes of their profession when AI can perform its most prestigious work; and researchers propose an AI adverse-event reporting system to detect harmful patterns in AI use before they develop into crises.

Flock Wants a Closely Surveilled World with No Exit

New Yorker contributor Brady Brickner-Wood takes a deep dive into how Flock Safety is building a privately operated, nationwide surveillance network that asks Americans to surrender any realistic ability to move anonymously in exchange for the promise of eliminating crime.

The surveillance of it all is conveniently ignored by people in power who laud Flock, and by Langley himself. “I’ve never met anyone who doesn’t want to be safer,” [Garrett Langley, the founder of Flock Safety] told NBC News, last month. “I’ve travelled the country, and I have not met that person yet.” Langley, it seems, has not been meeting the right people. In recent months, Flock cameras and the larger A.L.P.R. industry have faced mounting backlash, as people around the country, and across the political spectrum, denounce the spread of surveillance infrastructure into public life as undemocratic.

Almost half of Americans oppose A.L.P.R.s, according to the Washington Post, and some of them have taken to damaging, stealing, and defacing Flock cameras at parks, street corners, and traffic stops. When one Georgia sheriff posted a video to Facebook offering a thousand-dollar reward for information about the destruction of a Flock camera in a public park—“it was to keep our kids safe,” he said, from “sex offenders”—commenters ridiculed him, and lauded the assailant as a hero. “Looks like a Constitutional wind blew it over,” one user wrote; “Kudos to the citizen that threw tea in the harbor!” another added. The crowdsourced-data app DeFlock has become a leader in the anti-A.L.P.R. movement, providing a nationwide map of Flock cameras, alongside legal and community resources. Langley called the project “terroristic,” a comment he later tried to walk back. But the public tides seem to have turned against Langley’s hollow rhetoric of safety, and his evocation of terrorism rang alarm bells for anyone fluent in post-9/11 propaganda-speak. Telling people that they need to compromise on privacy is a polite way of framing a nonconsensual takeover; calling a community opposition movement “terroristic,” however, gives the game away. In her 2019 book, “The Age of Surveillance Capitalism,” Shoshana Zuboff diagnosed this tension, showing how concessions around privacy expose the paradox within Big Tech’s continual promise of progress: “The precise moment at which our needs are met is also the precise moment at which our lives are plundered for behavioral data, and all for the sake of others’ gain. The result is a perverse amalgam of empowerment inextricably layered with diminishment.”

Even leaders in the country’s most reliably red states, heeding the cries of their constituents, are now cutting ties with Flock. No one wants to get spied on, it seems, even Blue Lives Matter conservatives. Late last month, the Texas governor, Greg Abbott, instructed state agencies to freeze spending on Flock cameras; a few days later, Florida’s governor, Ron DeSantis, revoked permits for A.L.P.R.s on state highways and ordered their removal. Republican officials in tough-on-crime states such as Indiana, Kansas, and Tennessee have also endorsed A.L.P.R. restrictions, invoking the public’s right to privacy amid growing evidence that the technology is being regularly abused. Police officers have been accused of exploiting Flock to stalk ex-girlfriends, track love interests, and monitor family members. A sheriff’s deputy in Texas searched a data pool of more than eighty thousand Flock cameras to incriminate a woman who had allegedly received a medication abortion. ICE has also tapped Flock data with the help of municipal police departments in order to hunt down and detain immigrants. Langley has distanced himself from such deployments of his product by performing the role of the detached inventor. He told Politico that he viewed his “job as building tools and then giving those which we elect democratically the control to decide how they’re implemented.” He “didn’t want to have” the “responsibility” of regulating his products and making difficult ethical determinations, as he admitted to the BBC.

Listen to Langley talk long enough, and a core contradiction in his messaging emerges: he moralizes about safety and societal progress, about how his company is primed to change the course of human history, yet, when faced with genuine concerns about how his product creates less safe conditions for citizens, by repressing their autonomy, freedom, and right to public space, he holds up his hands, shakes his head, and pretends that none of that is any of his business. Flock cameras are meant to be a munificent gift from him to us. All he’s asking for in return is everything.