Toll Transponders Track You Beyond Toll Plazas

What a Toll Transponder Actually Broadcasts

The transponder on your windshield is not a payment device. It is a small radio beacon. When you drive under a reader at a toll plaza, the reader interrogates the transponder at 5.8 GHz or 915 MHz (depending on the regional standard), the transponder replies with its unique ID, and the reader's backend logs the timestamp, the lane, and your ID against the toll authority's billing system.^[3] That same handshake happens every time you pass any reader in the network, whether it is at a toll booth, a parking garage entry, or a "free flow" gantry on an open-road tolling stretch.

E-ZPass started as a 1993 pilot on the New York State Thruway and the Garden State Parkway, expanded into the E-ZPass InterAgency Group in 1995, and rolled out across the Northeast and Midwest through the late 1990s and 2000s.^[3] The same interoperable model now spans 25 agencies in 15 states, with parallel state-level systems (FasTrak in California, SunPass in Florida, TxTag in Texas, I-PASS in Illinois) all using compatible transponders on the same reader frequency. If your windshield sticker is any of those five brands, your unique ID is being read by a national network of tolling authorities.

The trip record is what gets assembled from those reads. It is not just the toll amount and the timestamp. It is a sequence of geolocated events tied to your transponder's ID, with start and end points, intermediate passes, time of day, lane, and (on most modern systems) the vehicle class. The format is standardized because the toll authorities need to settle payments with each other across state lines, which means the trip records are interoperable too.

Who Reads the Trip Records

The trip records are not a private database held by the tolling authority on your account. There are four documented readers.

1. The tolling authority itself. The agency you set the account up with holds the records for billing and audit. Retention policies vary by state; some authorities have published windows of three to seven years for the underlying trip data, longer for the billing summary.

2. Other tolling authorities. Because the systems are interoperable, the New York E-ZPass records for a Florida driver's trip across the Tappan Zee are settled between the two authorities' back offices. The records are shared between agencies on a routine billing basis.

3. Private contractors. Outside-the-toll-road-network readers, including parking garage readers and the open-road tolling gantries that private concessionaires operate, generate trip records that flow to a private operator rather than a public toll authority. The 2017 reporting on this beat documented per-trip records being resold to commercial buyers, including debt collectors and insurance companies. The original primary source for that reporting was a 2017 New York Times investigation that is no longer publicly accessible.

4. Law enforcement and immigration enforcement. Police agencies issue administrative subpoenas to tolling authorities for trip records, typically in support of an investigation rather than under a warrant. The standard that applies to those subpoenas, and whether Carpenter v. United States requires a warrant for historical trip records the way it does for cell-site location information, is the live legal question as of 2026.^[1]

The Electronic Frontier Foundation has been pressing the civil-liberties side of this question on plate-reader data since 2012, when it launched the Street-Level Surveillance project.^[2] EFF's broader argument is that automated tracking systems, whether they are toll transponders, fixed ALPR cameras, or mobile ALPR units mounted on patrol cars, generate the same kind of persistent location record, and the same constitutional question applies to all of them.

How Broken the Records Actually Are

The same problem that affects ALPR misreads affects toll transponder reads. A reader at a toll plaza can misfire, attribute a read to the wrong lane, log the wrong timestamp, or associate your transponder with a vehicle that is not yours. EFF's November 2024 piece on ALPR errors documented the pattern at the plate-reader layer: 1-in-10 plate reads are estimated to misread the state of issue, and other misreads (a '3' for a '7', an 'H' for an 'M', a '2' for a '7') are common enough that innocent drivers get stopped at gunpoint and detained on the strength of a bad read.^[2] The Aurora, Colorado case (Brittney Gilliam, August 2020, $1.9 million settlement) and the San Francisco case (Denise Green, March 2009, $495,000 settlement) both came from plate-reader errors, and both involved armed police response to a fabricated stolen-vehicle alert.^[2] Toll transponder records feed the same downstream enforcement pipelines, so the same error modes propagate.

The deeper issue is police over-reliance on the automated signal. EFF documents officers arriving at a suspect's house, handcuffing the occupants, and impounding the vehicle without checking the obvious physical signal (a fog light missing on the suspect vehicle, plate character mismatches) that would have caught the error in seconds.^[2] The same dynamic applies to toll-records subpoena practice: once the trip record is in front of an investigator, the question shifts from "is this right?" to "what does this person have to say about it?"

The Legal Landscape: Carpenter and Beyond

The legal foundation for any Fourth-Amendment challenge to toll-records access is Carpenter v. United States, the 5-4 Supreme Court ruling from June 22, 2018.^[1] The Court held that the government violates the Fourth Amendment when it accesses historical cell-site location information (CSLI) from a cellphone carrier without a warrant. The Court explicitly limited the so-called third-party doctrine, the legal principle that you have no reasonable expectation of privacy in records you voluntarily hand to a third party (your bank, your phone company, your email provider), at least for the precise category of records in Carpenter.

The Carpenter majority opinion was narrowly tailored. Chief Justice Roberts wrote the majority and was careful to limit the ruling to CSLI specifically, to avoid sweeping implications for other surveillance records. The dissent (Justice Kennedy, joined by Thomas and Alito, with Gorsuch in part) argued the Court was opening a door that could not be closed. The result is a ruling that everyone agrees is foundational and nobody agrees on the boundaries of.

The toll-records cases sit in that boundary zone. The argument for toll records being covered by Carpenter is structural: a toll transponder generates a persistent, retroactive, comprehensive record of the vehicle's location over time, exactly the category of record the Carpenter Court was worried about. The argument against is the formal third-party-doctrine point: you voluntarily gave the transponder to the toll authority and accepted the toll authority's records as a third party. Lower federal courts have been working through that question in the years since 2018, and the toll-records application has not been definitively resolved at the Supreme Court level as of 2026.

EFF has been on the Fourth-Amendment side of this question for the duration of the post-Carpenter litigation, both directly through amicus briefs and indirectly through its Street-Level Surveillance project and its ongoing plate-reader work.^[4] The question of whether a warrant is required for toll records specifically remains an open fight.

How ALPR Networks Make It Worse

The toll transponder is one of several automated tracking systems that share the same roadway. The other is the automated license plate reader (ALPR), the network of fixed and mobile cameras that read every plate that passes and upload it to a centralized hot-list database.^[5] ALPR systems read plates that have nothing to do with tolling, on roads that have nothing to do with tolling, and the resulting database covers most of the vehicle miles traveled in any major U.S. metro area.

When ALPR and toll-records data end up in the same investigative pipeline, the two systems reinforce each other. A plate-reader hit can locate a vehicle, a toll-records subpoena can fill in the gap (where the ALPR coverage is sparse), and the resulting picture is a near-complete reconstruction of where a vehicle has been. The privacy implication is the sum, not the parts. Either system alone is a tracking system. Both together, with shared access by the same agencies, is a surveillance system with nationwide coverage and no opt-out.

Flock Safety operates one of the largest ALPR networks in the United States and runs roughly 20 billion plate reads a month across its customer base. The same network has documented sharing with U.S. Immigration and Customs Enforcement, and at least one federal court case has produced a 1.6 million-record disclosure figure for California plate data shared with federal authorities.

What You Can Actually Do

There is no clean way to opt out of the toll-transponder network without changing how you drive. The practical options are limited.

• If you do not use toll roads: remove the transponder from your windshield. Most authorities will close the account on request. The transponder only reads at readers you pass, and it does not read at all if it is not mounted. (Some newer authorities charge a small monthly fee for an unused account, which is its own kind of pressure; check the terms before you set one up.)
• If you use toll roads occasionally: most toll authorities now offer a pay-by-plate option, where a camera reads your plate and bills the registered owner. The camera still logs the read. The pay-by-plate option does not give you anonymity; it gives you the option of not having a transponder in the windshield. The trade-off is higher per-trip fees and a slower billing cycle.
• If you want to minimize the data: ask the toll authority for its records-retention policy in writing. Several authorities have published three-to-seven-year retention windows for the underlying trip data. The retention window is the upper bound on what a subpoena can pull. Some authorities will purge on request after the account is closed.
• If you are subject to a subpoena: the toll-records subpoena question is the active legal fight. If you receive notice that a tolling authority has produced your records in response to a subpoena, the EFF maintains a Street-Level Surveillance hub with the legal-resources links to challenge the production. The Carpenter precedent is the legal foundation; the application to toll records specifically has not reached the Supreme Court.
• If you write about this: cite the original 2017 NYT piece by name (even though the URL is no longer accessible), cite the EFF Street-Level Surveillance hub for the civil-liberties framing, and cite Carpenter v. United States (2018) for the legal foundation. The provenance matters because the case law is in flux and a sloppy citation will be challenged in the comments.

Sources

1. Wikipedia: "Carpenter v. United States" (585 U.S. 296 (2018), the June 22, 2018 5-4 Supreme Court ruling that government acquisition of historical cell-site location information is a search under the Fourth Amendment and requires a warrant, the third-party-doctrine limitation, the narrowly-tailored majority opinion by Chief Justice Roberts, and the Kennedy-Thomas-Alito dissent arguing the ruling opened a door that could not be closed)
2. EFF Deeplinks: "The Human Toll of ALPR Errors" (Gowri Nayar, November 1, 2024, the EFF legal team's broader plate-reader civil-liberties work, the wrongful-detention cases Brittney Gilliam in Aurora Colorado August 2020 and Denise Green in San Francisco March 2009, the Jason Burkleo Atherton California and Jaclynn Gonzales Espanola New Mexico cases, the Brian Hofer Contra Costa 2019 case, the Isoke Robinson Detroit case, the 1-in-10 state-misread rate, the $1.9M Aurora settlement, the $495,000 San Francisco settlement, the $49,500 Hofer settlement, and EFF's Street-Level Surveillance project launched in 2012)
3. Wikipedia: "E-ZPass" (the 1993 New York State Thruway and Garden State Parkway pilot, the 1995 E-ZPass InterAgency Group formation, the multi-state rollout from 1997 to 2017, the 25-agency 15-state membership, and the interoperable systems FasTrak in California, SunPass in Florida, TxTag in Texas, and I-PASS in Illinois)
4. EFF: "Automated License Plate Readers (ALPRs)" topic page (the EFF's standing institutional position on plate-reader systems, the nationwide-deployment data, the documented sharing with immigration enforcement, the link out to the Street-Level Surveillance project, and the legal-resources landing for individuals subject to plate-reader subpoenas)
5. Wikipedia: "Automatic number-plate recognition" (the system mechanics, the optical-character-recognition pipeline, the hot-list lookup against vehicles of interest, the per-camera per-day read volumes, and the vendor landscape including Vigilant Solutions, Flock Safety, Motorola Solutions, and Rekor Systems)