Blog · Analysis · Last reviewed June 24, 2026

The Supervision App Becomes the Pocket Probation Officer

Smartphone supervision apps promise a less visible alternative to ankle monitors. They also turn the phone into a biometric, locational, behavioral checkpoint for people living under court, parole, probation, or immigration control.

For this essay, pocket probation means a legal-supervision workflow in which a personal or issued mobile device becomes the instrument for identity proof, location evidence, reminders, reporting, exception handling, and sanction risk. The governed object is the whole chain: order, device requirement, permission prompt, biometric or location sample, alert, human review, sanction decision, and record disclosure. The test is not whether the app feels gentler than jail. The test is whether it reduces custody or merely makes custody portable.

From Ankle to Phone

The old image of electronic monitoring is the ankle bracelet: visible, stigmatizing, physically attached, and hard to forget. The new image is quieter. A person downloads an app, grants permissions, checks in with a face or fingerprint, shares location, receives reminders, uploads evidence, answers questions, and waits for the system to decide whether the interaction counts as compliance.

A supervision app is not an ordinary compliance app. It is software used to enforce a legal or administrative condition of release: identity verification, location confirmation, schedule compliance, treatment attendance, sobriety testing, document exchange, or communication with a supervising officer or contractor. It becomes a pocket probation officer when a missed prompt, failed face match, GPS drift point, dead battery, or unsupported phone can trigger officer attention, sanctions, detention, removal risk, or a return to court.

A supervision app event is not the same thing as a violation. It is a recorded interaction or exception: prompt issued, check-in completed, location captured, biometric verification passed or failed, message sent, data delayed, alert generated, or support ticket opened. A violation finding should require a separate legal step that names the rule, the authority, the evidence, the person's explanation, and the reviewer who decided that the event means noncompliance.

That change matters because community supervision is already a massive administrative layer. The Bureau of Justice Statistics reported that at yearend 2024, about 3,681,900 adults in the United States were under community supervision, including 3,030,500 on probation and 663,800 on parole. BJS also warns that 2023 and 2024 probation counts are affected by expanded survey coverage, so the point is scale rather than a simple year-over-year trend. Most people under supervision are not on location monitoring. But even a small technological shift inside a population that large can create a major governance system.

The official language emphasizes alternatives. The federal courts describe location monitoring as a court-ordered alternative to detention or imprisonment, allowing supervised activities such as work, school, treatment, religious activity, court obligations, and approved appointments. That can be materially better than jail. A person who keeps housing, employment, family contact, and treatment access has a real chance at stability.

But an alternative to incarceration is not automatically freedom. It may be a different form of custody, distributed through devices, schedules, alerts, vendors, fees, app stores, and case-management dashboards. The jail cell recedes. The checkpoint follows.

Current Context

As of June 24, 2026, smartphone supervision sits across several overlapping regimes. Federal probation and pretrial services use virtual mobile applications as one location-monitoring option. State, local, parole, probation, pretrial, treatment, and specialty-court programs buy smartphone reporting, location, alcohol-testing, reminder, and case-management products from private vendors. ICE's Alternatives to Detention program uses SmartLINK, body-worn GPS, wrist-worn devices, and telephonic reporting in the Intensive Supervision Appearance Program.

The current public count is clearer for ICE than for the whole country. ICE's FY 2026 detention-management spreadsheet lists 180,701 active ATD participants as of April 4, 2026, including 131,643 assigned to SmartLINK, 46,302 to ankle monitors, 2,707 to wrist-worn devices, and 49 to dual technology. That is not a national count of smartphone supervision. It is a snapshot of one federal immigration program, but it shows the order of magnitude and the degree to which the app has become the dominant ATD technology.

The strongest current ICE source for the app's claimed technical boundary is the Department of Homeland Security's ATD privacy-impact assessment, not a product slogan. DHS describes the Monitoring App, also known as BI SmartLINK, as a check-in tool that may use facial verification and location data at login or check-in, while stating that it is not designed for continuous location tracking and that it should not access contacts, photos, text messages, or other personal phone content. That official description is still not independent proof of lived experience, error rates, or whether the person can meaningfully refuse; it is the agency's stated collection and use boundary.

The criminal legal side remains harder to count. Pew's 2016 survey estimated more than 125,000 people on GPS or radio-frequency electronic tracking in 2015. Vera's 2024 report estimated 254,700 adults on electronic monitoring in 2021 across criminal legal and civil immigration systems, with ICE monitoring rising sharply in 2022 and pushing the total close to half a million. Those are careful estimates, not a standing public census. The public still lacks routine national reporting by legal status, technology type, vendor, fee burden, violation type, sanction, malfunction, and outcome.

The Status Boundary

The phrase "community supervision" hides important legal differences. Pretrial release, probation, parole, supervised release, diversion, treatment-court monitoring, and immigration ATD are not the same authority. They involve different decisionmakers, burdens of proof, constitutional settings, counsel access, records, remedies, and consequences. A missed app check-in may mean one thing in a federal location-monitoring program, another in a county pretrial program, another in parole, and another in immigration enforcement.

That boundary matters because app design tends to flatten legal status into a common dashboard vocabulary: participant, assignment, check-in, alert, violation, escalation, closure. A vendor feature built for "supervision" can migrate across regimes faster than public rules can follow. Governance has to travel with the legal authority, not with the software category. The same face match, GPS point, or missed notification should not silently acquire the same meaning everywhere.

Every consequential app record should therefore carry a status label. A pretrial location exception is not a probation technical violation. An immigration ATD prompt is not a criminal sentence condition. A treatment-court breath test is not the same record as a parole geofence alert. Without that label, downstream reviewers can inherit the force of the system without seeing the source of authority.

This is why the supervision app belongs beside notice and appeal, AI audit trails, and data minimization even when no generative AI is involved. The app turns administrative power into records. Those records need purpose limits, review paths, and a clear account of who is allowed to act on them.

What the App Does

The federal judiciary recognizes virtual mobile applications as one of the technologies available for location monitoring. Its public materials describe a model in which the participant uses a smartphone or tablet, GPS location services, identity technology such as facial recognition or fingerprint verification, and a password to verify presence at a location. The same federal guide distinguishes these apps from continuous 24/7 GPS ankle tracking: virtual mobile app supervision is framed as periodic spot-checking location sampling, generally suited for lower-risk, nonviolent people under supervision.

The National Institute of Justice's Criminal Justice Testing and Evaluation Consortium summarizes the broader category: smartphone supervision apps may support remote reporting, treatment delivery, location tracking, calendars and reminders, goal setting, substance testing, accountability, behavior change, and case-management objectives. The promise is not only surveillance. It is supervision made portable.

The operational unit is the workflow, not the icon on the phone. A prompt is sent. The participant responds. The app records time, identity signal, location signal, device or network context where available, and completion status. A vendor system or agency dashboard turns that event into a record. An officer, case manager, or court may later treat the record as evidence of compliance, noncompliance, risk, or technical failure. The app is therefore both a sensor and a clerk.

The useful evidence unit is the event packet: legal condition, prompt text, scheduled window, device or app version where available, permission state, timestamp, location source and confidence where available, biometric-verification result, delay or outage marker, participant response, officer action, vendor support record, and final disposition. Without that packet, a dashboard status can outrun the evidence it claims to summarize.

That packet should also separate system fact from legal meaning. A location point says where a device reported itself. A biometric result says whether a threshold was met. An alert says the system routed attention. None of those facts should by itself answer whether the person violated a condition.

That is precisely why the interface is powerful. An app can look less punitive than an ankle monitor while reaching more deeply into everyday life. It may ask for the camera, microphone, location, notifications, device identifiers, Bluetooth connection to a breathalyzer, or access patterns that make the phone part of the supervision apparatus. It may become the place where the person proves presence, sobriety, responsiveness, schedule discipline, and identity. In the terms used elsewhere on this site, it is a public-power interface, not merely a product feature; see AI in Government and Public Services and Human Oversight of AI Systems for the broader governance pattern.

The phone is not just a sensor. It is a social organ. It holds work messages, child-care coordination, banking, maps, intimate communication, job applications, medical portals, photographs, calendar reminders, emergency calls, and ordinary boredom. Turning that device into a court-facing checkpoint changes the meaning of carrying it.

Scale Without Public Memory

The public record on electronic monitoring is weaker than the power of the system deserves. BJS counts probation and parole. ICE publishes detention and ATD spreadsheets. Federal courts publish guidance and approximate location-monitoring counts. Researchers and advocacy organizations assemble broader estimates. But no simple public table tells a resident how many people in the United States are currently under smartphone supervision, what software they use, who pays, what data is retained, how often alerts are wrong, or how often app events become sanctions.

That absence is itself a governance fact. The country can expand a supervision technology faster than it can publicly describe the system it has built. Counting is harder because "monitoring" now spans ankle GPS, radio-frequency home detention, wrist-worn devices, voice check-ins, face-matching smartphone apps, alcohol-testing peripherals, treatment portals, appointment reminders, and case-management messaging. Some systems continuously track. Others spot-check. Some are court-ordered. Some are immigration conditions. Some are vendor-operated. Some are paid by the person monitored. The public debate often collapses all of that into the phrase "alternative to detention."

The pocket probation officer is no longer hypothetical at scale, and the clearest current example is still immigration supervision. ICE's archived 2024 ATD page described SmartLINK as using biometric facial comparison, participant location confirmation during check-in, reminders, document upload, messaging, and community-resource information. It also said SmartLINK may obtain a single GPS point at login or scheduled check-in and that most ATD-ISAP participants were assigned to either SmartLINK on a personal phone or an issued SmartLINK device. The page is now archived, so it should not be treated as a live policy statement. But combined with ICE's FY 2026 spreadsheet, it documents how app-based supervision became the dominant technology in ATD.

Smartphone supervision worsens the memory problem because the boundary of "monitoring" becomes harder to see. Is a check-in app electronic monitoring? What if it only collects location when opened? What if it also sends reminders, records treatment attendance, verifies face, supports alcohol testing, or lets officers message participants? What if a person must keep notifications on, keep a data plan active, and respond within a narrow window? A dashboard may treat these as features. For the person under supervision, they are conditions of liberty.

The site has already treated the real-time crime center as a city dashboard, the face as a ticket, and the police report as institutional memory. The supervision app brings the same logic to the person who is not in a control room and cannot opt out. It makes compliance measurable at the scale of daily movement.

The Private Device Problem

Smartphone monitoring often depends on the participant's own device. The federal courts' chapter on location monitoring notes that virtual monitoring requires a mobile device with a data plan and enabled location services. That sounds administratively efficient. It also transfers part of the infrastructure burden onto the person being monitored.

A person may need a compatible phone, enough storage, battery life, camera access, data service, notifications, and the ability to keep the device charged and functioning. A missed check-in may reflect disobedience. It may also reflect a dead battery, broken screen, poor signal, app crash, camera failure, job-site restriction, disability, language barrier, or lack of money for service. The interface can translate poverty and technical fragility into suspicion.

The fallback cannot be an afterthought. If a court or agency makes a phone part of liberty, the order should say what happens when the phone is lost, stolen, unaffordable, incompatible, confiscated at work, inaccessible because of disability, or unusable because a shelter, hospital, school, courthouse, or job site restricts devices. Otherwise the state has delegated a custody condition to a consumer object the person may not control.

Bring-your-own-device supervision also turns ordinary privacy choices into compliance hazards. A participant may be unable to deny location permission, disable notifications, use a privacy-preserving operating-system setting, uninstall a tracking library, replace a broken phone slowly, or refuse a biometric workflow without risking an adverse report. That is the digital poorhouse pattern in miniature: public power depends on private infrastructure, and the cost of failure lands on the person least able to absorb it.

Accessibility is part of the same problem. For state and local public entities, DOJ's Title II web and mobile-app rule treats mobile apps used to provide government services, programs, and activities as accessibility infrastructure, including apps made available through contractual arrangements. The April 2026 interim final rule extended compliance dates, but it also left existing Title II obligations in place. A supervision condition should therefore include effective communication, reasonable modification, and a nonpunitive fallback before the app fails a disabled user.

The 2022 USENIX Security paper on electronic monitoring smartphone apps analyzed sixteen apps used as alternatives to criminal and civil detention. The researchers found privacy and security concerns around permissions, third-party libraries and trackers, data flows, usability, and malfunction reports. EFF's summary of the study emphasized a basic problem: many users do not meaningfully consent to these apps in the ordinary consumer sense. They install because liberty depends on installation.

That coercive context should change the privacy analysis. A social app that overreaches is already a problem. A supervision app that overreaches sits inside state power. The user cannot easily delete it, deny a permission, refuse an update, or switch providers without risking noncompliance. Personal-device supervision therefore belongs inside the same institutional obligations described in Privacy and Data, Vendor and Platform Governance, and AI Audits and Third-Party Assurance: minimization, security, retention limits, procurement transparency, independent review, and meaningful recourse.

Location doctrine is not a product-design footnote either. In Carpenter v. United States, the Supreme Court recognized that historical cell-site location records can reveal the privacies of life. Supervision apps are different because the person is under a release, sentence, court, or immigration condition. But the privacy point remains: location traces and phone-dependent compliance are not neutral administrative residue.

Violations Become Interface Events

The federal courts are careful about the limits of location monitoring. Their guidance says location data is not necessarily sufficient to determine compliant or noncompliant behavior and must be corroborated through other supervision strategies, such as community visits and collateral contacts. That sentence should be treated as a constitutional warning for the interface.

Location is not conduct. A GPS point is not intent. A missed notification is not flight. A failed facial comparison is not deception. A delay is not proof of danger. A device signal can be evidence, but evidence must be interpreted inside a human and legal context. The federal guide's treatment of GPS drift points shows why: atmospheric conditions, satellite orbit, humidity, and signal reflection off buildings can produce apparent location anomalies that require investigation rather than automatic punishment.

The risk is that the app turns ambiguity into workflow. The system emits an alert. The officer sees an exception. The participant becomes a case requiring response. The dashboard compresses context into a status. The person must explain why the machine did not record compliance in the expected form. In that loop, the burden of interpretation shifts downward.

Due process therefore needs an error channel, not only an alert channel. Participants, counsel, case managers, and courts should be able to see the relevant record: requested check-in, timestamp, device state where available, GPS confidence or drift indicators, biometric failure reason if available, officer response, vendor communication, and the path for contesting the event before liberty is restricted.

The review interface should preserve ambiguity instead of erasing it. If the only durable fields are "missed," "late," "out of zone," or "failed identity," the record has already converted a technical classification into a legal posture. A defensible record should retain uncertainty markers, known outages, support tickets, participant messages, and any human decision not to escalate.

No app-only event should become an automatic sanction. The institution should have to separate technical failure, inability, ambiguity, emergency, and willful violation before escalating. It should also preserve the evidence needed for the participant to challenge the event. A notice that says only "missed check-in" is not enough when the underlying cause may be connectivity, accessibility, vendor error, device failure, or a confusing instruction.

This is the same administrative pattern described in The AI Detector Becomes the Discipline Machine and The Adverse Action Notice Becomes the Explanation Interface. The score, flag, or notice is not the whole decision. But once it enters a disciplinary workflow, it can become the event around which the person must organize a defense.

Fees and Vendors

Electronic monitoring is also a market. Vendors supply devices, apps, dashboards, call centers, alcohol testing systems, data infrastructure, support, and analytics. Courts and agencies may pay. In some jurisdictions, the monitored person pays.

In July 2024, U.S. Senators Elizabeth Warren, Ron Wyden, and Ed Markey sent letters to electronic monitoring companies raising concerns about fees, privacy, competition, and consumer protection. The letter cited user fees often ranging from $5 to $30 per day, additional charges for mandatory services, and cases where annual costs can exceed $3,000 for a person on electronic monitoring. The letter is not neutral research, but it identifies the institutional pressure clearly: when liberty is mediated by a vendor system, the price of compliance can become part of the punishment.

Vendor structure matters even when the participant is not charged directly. GAO reported in 2022 that ICE used a $2.2 billion contract to administer ATD and needed better performance assessment and contract oversight. ICE's FY 2026 spreadsheet says the listed ATD technology costs do not include other associated contract and case-management costs. That distinction matters because public debate often compares a low daily technology cost to detention while leaving vendor oversight, case-management labor, support, errors, app infrastructure, and participant burden outside the headline number.

Fees also change incentives. A system that charges per day can normalize longer supervision. A jurisdiction that offloads costs to monitored people can expand monitoring without making the public budget reflect the full burden. A vendor that sells lower-cost smartphone check-ins can help reduce reliance on more restrictive hardware, but it can also make it easier to monitor people who otherwise would not have been electronically monitored at all.

That is net-widening: a tool introduced as an alternative to jail becomes an additional layer on people who might otherwise have had less surveillance. The humane version replaces confinement. The punitive version makes supervision cheaper, wider, and more continuous.

Failure Modes

Net-widening is the first failure mode, but not the only one. A jurisdiction can sincerely reduce jail admissions for some people while also placing new groups under app-based reporting because the marginal cost looks low and the device is already in the person's pocket. The right denominator is therefore not app versus jail in the abstract. It is app versus the least restrictive actual alternative for that person.

The second failure mode is event inflation. Prompts, late responses, GPS drift, face-match failures, support tickets, and battery gaps can accumulate as if they are behavior. If the dashboard preserves only the exception and not the explanation, the record becomes more punitive each time it is reused.

The third is status laundering. A feature built for immigration ATD, pretrial release, probation, parole, or treatment compliance can cross into another setting with the same labels but different law. "Participant," "alert," and "violation" can look portable even when the underlying authority is not.

The fourth is device poverty becoming legal risk. A person with an old phone, unstable data plan, inaccessible interface, shared device, low battery, or unsafe housing situation may look less compliant than a person with reliable infrastructure. That is a social condition translated into a supervision signal.

The fifth is vendor opacity and secondary use. If contracts, SDKs, analytics, retention schedules, model changes, support logs, breach notices, and deletion rules are not inspectable, the public cannot know whether the app is doing only the task the court or agency ordered. A narrow check-in tool can quietly become a broader location, identity, and behavior record.

The Governance Standard

A serious supervision-app regime should treat the app as carceral infrastructure, not as ordinary consumer software.

First, use should be tied to a specific risk and reviewed regularly. A person should not be placed on smartphone monitoring because it is available, cheap, or administratively convenient. The court or agency should state the risk, the required feature, the least restrictive alternative, the duration, and the review date.

Second, app data should be minimized. The system should collect only what is necessary for the supervision condition: not broad device data, not continuous location when periodic verification is enough, not unrelated behavioral traces, not hidden third-party analytics, and not indefinite archives.

Third, technical failure should not automatically become violation evidence. Dead batteries, app crashes, poor connectivity, inaccessible interfaces, biometric mismatch, device incompatibility, and affordability problems need documented review before they become sanctions.

Fourth, participants need plain-language receipts and contestability. A person should be able to see what check-ins were requested, what check-ins were recorded, what location points were captured, what alerts were generated, what data was shared, and how to contest an error before the event becomes sanction evidence.

Fifth, vendors should be auditable. Contracts should cover data retention, security, third-party SDKs, biometric performance, breach notice, algorithmic or matching changes, accessibility, support response, deletion, and public-record access where lawful. Procurement terms should not let a vendor hide the evidence trail needed by courts, defense counsel, public defenders, participants, auditors, or journalists.

Sixth, fees should not be a hidden sentence. If monitoring is court-ordered, the cost structure should be visible to the court, the participant, defense counsel, and the public. Inability to pay should not convert an alternative to jail into a debt trap.

Seventh, aggregate reporting should be mandatory. Jurisdictions should publish counts by legal status, technology type, duration, fees, violations, sanctions, malfunctions, demographic distribution, vendor, and outcome. A system that can locate people should be able to locate its own accountability record.

Eighth, independent impact assessment should precede expansion. Before an agency moves from ankle monitors to smartphone apps, from check-ins to continuous tracking, or from human scheduling to automated alerts, it should publish an impact assessment, retention schedule, access policy, appeal path, and sunset review. The standard is close to an algorithmic impact assessment even when the app is not marketed as artificial intelligence, because the system still turns data into institutional action.

Ninth, legal status should stay visible. The record should identify whether the person is on pretrial release, probation, parole, supervised release, immigration ATD, diversion, or another authority. The app should not let one generic vendor category obscure the legal route by which liberty is restricted.

Tenth, accessibility and fallback must be part of the order. Language access, disability accommodation, camera alternatives, non-smartphone options, device replacement, charging access, signal gaps, job-site restrictions, and emergency procedures should be written into the supervision plan before the first failed prompt.

Eleventh, data sharing should be bounded. Location, biometric, communication, device, and compliance records should not flow to unrelated policing, immigration, commercial analytics, marketing, or model-training uses without explicit lawful authority, documented purpose, and public accountability.

Twelfth, incident reporting should include false violations. Agencies should track app outages, biometric failures, drift alerts, missed-notification disputes, support delays, vendor errors, sanctions later reversed, and detention or revocation events that began as app alerts.

Thirteenth, distinguish alert, violation, and sanction. The record should show when a system event occurred, when a human found a rule violation, and when a court, agency, or officer imposed a consequence. Collapsing those moments makes the app look more authoritative than it is.

Fourteenth, prove custody reduction. Agencies should report how many people were placed on the app instead of jail, detention, or an ankle monitor; how many were placed on the app who otherwise would have had no electronic monitoring; and how many were escalated because of app-generated events. Without that accounting, "alternative" can become net-widening.

Fifteenth, protect defense and participant access. If app records can support detention, revocation, removal, or a warrant, affected people and counsel need timely access to the underlying event packet, vendor notes, relevant logs, and correction path, subject to narrow safety limits.

Sixteenth, preserve the source-to-sanction chain. A reviewer should be able to move backward from sanction to violation finding, from violation finding to alert, from alert to event packet, and from event packet to the underlying condition and human reviewer. If the chain cannot be reconstructed, the app record is not strong enough to restrict liberty.

Seventeenth, test accessibility before deployment. Required apps should be tested with assistive technologies, language access, low-bandwidth conditions, older devices, shared devices, and non-smartphone alternatives. A person should not have to generate a failed check-in to prove that the compliance interface did not work for them.

What This Changes

The supervision app is a high-control interface because it makes freedom conditional on a device interaction.

The person is outside the jail, but the institution is inside the pocket. It asks for presence, face, location, response, schedule, sobriety, and proof. It speaks in reminders and alerts rather than bars and locks. It can be less cruel than confinement. It can also make the world feel like a distributed checkpoint.

The deeper issue is model-mediated and dashboard-mediated knowledge. A person's life becomes visible to an officer through pings, exceptions, maps, check-ins, risk categories, and vendor records. The officer may still exercise judgment, and the best systems require corroboration. But the interface shapes what counts as the first version of reality. It says where the person was, whether the face matched, whether the response was late, whether the route was strange, whether the rule was touched.

That first version matters. Institutions often treat machine-readable traces as cleaner than human explanation. People under supervision then learn to live for the trace: keep the battery charged, stay near signal, answer immediately, perform identity for the camera, preserve proof against the app, and organize life around the possibility of a false alert.

The practical discipline is not to romanticize old punishment or reject every alternative. Jail is violence. Community supervision can preserve work, family, treatment, and life outside a cage. But an alternative worthy of the name must reduce custody rather than miniaturize it.

The phone should not become a portable cell door. If supervision must use an app, the app should be narrow, contestable, temporary, auditable, affordable, accessible, and subordinate to human judgment. Otherwise pocket probation becomes another way for institutions to call control convenience.

Source Discipline

The source stack needs separation. BJS gives national probation and parole population counts, not a count of smartphone supervision. ICE gives a current federal immigration ATD spreadsheet, not a national criminal-legal census. U.S. Courts materials describe federal location-monitoring policy and technology categories, not every state or local practice. NIJ and CJTEC describe the smartphone-supervision technology landscape and implementation considerations, not a guarantee of outcomes.

DHS's ATD privacy-impact assessment is an official description of ICE's program design, stated safeguards, collection logic, and privacy risks. It is not an independent audit of SmartLINK, not a participant-experience study, and not proof that every deployment follows the stated boundary. Its proper use is narrower: to identify what the agency says the app may collect, when location data is transmitted, what personal-phone access is prohibited, what retention rules apply, and where the agency itself acknowledges risk.

The broader electronic-monitoring numbers from Pew and Vera are estimates built from available public records, not a live registry. GAO's ATD report is an oversight source about federal contract performance and program assessment. The USENIX paper is a technical and human-centered security study of sixteen Android apps. The senators' 2024 letter is congressional oversight advocacy; it is useful for fees and market questions, but its claims should be read with its posture in mind.

Carpenter and the ADA mobile-app sources supply legal risk frames, not app-specific supervision rules. Carpenter is about historical cell-site location information, not probation software. DOJ's Title II web and mobile-app rule applies to state and local public entities and their covered arrangements, not every federal or private deployment. They matter here because required mobile check-ins convert location privacy and accessibility into conditions of liberty.

That discipline keeps the core claim narrow: smartphone supervision is real, consequential, and undercounted. The verified public record supports concern about scale, privacy, security, fees, ambiguity, and oversight. It does not support treating every app check-in as continuous tracking, every ATD participant as a criminal defendant, or every "alternative" as either freedom or jail. The governance question is more precise: under what authority, with what data, at what cost, with what recourse, and with what evidence that custody actually decreased?

Sources

Return to Blog