General Motors Pays $220 Million Over Secretly Selling Driver Location Data

General Motors has agreed to pay $220 million to settle a sprawling multistate investigation into the covert harvesting and monetization of driving records, resolving claims that the automaker spent years capturing high-frequency trip metrics and quietly feeding them to the insurance industry.

Announced by a bipartisan coalition of 36 state attorneys general and the District of Columbia, the settlement closes one of the most aggressive regulatory inquiries ever conducted into connected-vehicle technology. Co-led by Oregon Attorney General Dan Rayfield and Washington Attorney General Nick Brown, state investigators revealed that GM converted millions of passenger vehicles into commercial tracking appliances, turning routine errands, commute times, and split-second braking events into commodities.

"People bought a car, not a tracking device," Rayfield said when announcing the resolution. "GM told drivers this program would help them drive more safely and promised it wouldn't share their data with insurers without permission. Our investigation found GM sold that data anyway, and some drivers paid for it with higher premiums. This settlement puts people back in control of their own information".

The monetary sanction—one of the largest privacy-related penalties ever leveled against an industrial manufacturer—punches a massive hole in Detroit’s multi-billion-dollar ambition to turn vehicles into software platforms. Behind the headline figure lies an intricate pipeline of embedded hardware, Controller Area Network (CAN) bus sniffing, dealership sales quotas, and actuarial arbitrage.

Between 2016 and mid-2024, GM enrolled an estimated eight million motorists into its OnStar Smart Driver program. For a large portion of those drivers, vehicle telemetry was systematically funneled through data brokers like LexisNexis Risk Solutions and Verisk Analytics directly to underwriters, sparking unprompted insurance rate spikes for motorists who had never been in an accident or received a citation.

The resolution requires GM to overhaul its vehicle data collection practices, establish binding technical boundaries for dealerships, and secure unbundled, affirmative express consent before pulling telematics off the road. More than an enforcement action, the legal record unsealed across dozens of state capitals exposes how the automotive industry constructed an invisible surveillance apparatus behind the dashboards of modern cars.

                     THE CONNECTED VEHICLE SURVEILLANCE PIPELINE
                                (2016 – 2024)

  +-----------------------+
  |  GM PASSENGER VEHICLE |
  |  (Chevy, GMC, Buick,  |
  |       Cadillac)       |
  +-----------+-----------+
              |
              | CAN Bus Telematics (High-Frequency Ingestion)
              | - Geolocation logged as often as every 3 seconds
              | - Speed, throttle position, hydraulic brake pressure
              | - Tri-axial accelerometer (lateral / longitudinal Gs)
              v
  +-----------------------+
  |  ONSTAR TELEMATICS    |
  |     CONTROL UNIT      | <--- Dealership Delivery Sign-up
  +-----------+-----------+      (Bundle-click enrollment & dark patterns)
              |
              | Encrypted 4G LTE/5G Cellular Uplink
              v
  +-----------------------+
  |  GM VEHICLE DATA      |
  |     PLATFORM          |
  +-----------+-----------+
              |
              | Commercial Data Broker Contracts
              v
  +--------------------------------------------------------+
  |                   DATA BROKER INGESTION                |
  |  +---------------------------+  +-------------------+  |
  |  | LexisNexis Risk Solutions |  | Verisk Analytics  |  |
  |  | (Telematics Exchange)     |  | (DriveInformed)   |  |
  |  +-------------+-------------+  +---------+---------+  |
  +----------------|--------------------------|------------+
                   |                          |
                   +-------------+------------+
                                 |
                                 | Consumer Reporting Profiles
                                 | (Driving Behavior Risk Scores)
                                 v
                 +-------------------------------+
                 |     AUTO INSURANCE INDUSTRY   |
                 |  - Policy premium hikes       |
                 |  - Coverage non-renewals      |
                 |  - Dynamic risk profiling     |
                 +-------------------------------+

The Hardware and Telematics Architecture Behind the Wheel

Most drivers assume that when a vehicle engine shuts off, the relationship between the machine and the manufacturer pauses until the next ignition cycle. In a modern GM vehicle, the architecture works under a fundamentally different operating logic.

The nervous system of every vehicle manufactured over the last three decades is the Controller Area Network, an internal communications protocol developed by Robert Bosch GmbH in the 1980s. In contemporary GM platforms—from a Chevrolet Silverado to a Cadillac Lyriq—the vehicle operates multiple high-speed and low-speed CAN buses, as well as Ethernet backbones. These digital corridors exchange millions of operational messages every second between specialized microcontrollers known as Electronic Control Units (ECUs).

 +-------------------------------------------------------------------+
 |                   INTERNAL VEHICLE CAN BUS TOPOLOGY               |
 +-------------------------------------------------------------------+
           |                     |                       |
           v                     v                       v
     [ Engine ECU ]       [ Braking ECU / ABS ]    [ Airbag / IMU ]
           |                     |                       |
           +----------+----------+-----------+-----------+
                      |                      |
                      v                      v
           [ Body Control Module ]  [ Wheel Speed Sensors ]
                      |
                      +-------------------+
                                          |
                                          v
                         +----------------------------------+
                         |  Telematics Control Unit (TCU)   |
                         |  - Gen10 / Gen11 OnStar System   |
                         |  - Dedicated Multi-Core SoC      |
                         |  - High-Precision GNSS Receiver  |
                         |  - Internal 3-Axis Gyro/Accel    |
                         |  - eSIM Cellular Baseband (LTE)  |
                         +----------------------------------+
                                          |
                                          | In-Vehicle Data Ingestion
                                          v
                         [ Local Flash Memory Ring Buffer ]
                                          |
                                          | Pushed via MQTT / REST API
                                          v
                            GM Cloud Telemetry Servers

Every foot tap on the accelerator, every fraction of pedal travel on the master brake cylinder, and every yaw change tracked by stability control sensors broadcasts unencrypted diagnostic messages across the CAN network.

Mounted behind the instrument cluster or concealed beneath the center console sits the Telematics Control Unit (TCU). In GM vehicles running Gen10 and Gen11 architectures, the OnStar TCU is not a passive receiver waiting for an emergency SOS signal; it is an active, edge-computing network node equipped with its own multi-core system-on-chip, independent flash memory, a cellular baseband modem with an embedded SIM (eSIM), and a dedicated satellite navigation receiver.

The investigation conducted by state attorneys general, echoing granular findings by the Federal Trade Commission, revealed that GM modified the operational profiles of these TCUs through back-end micro-tasks. When configured for Smart Driver telemetry, the unit ran persistent background software routines designed to listen to the CAN bus traffic and sample vehicle state metrics with continuous precision.

Federal regulators discovered that in millions of Gen10-equipped vehicles, GM executed background scripts that sampled and logged precise GPS coordinates as frequently as every three seconds.

This frequency was not reserved for active turn-by-turn navigation sessions. The module was logging coordinates during ordinary local trips, generating a breadcrumb trail dense enough to reconstruct precise travel vectors, driving lanes, intersections crossed, and curb parking locations.

Raw Telematics Packet (Generated every 3 seconds by TCU):
----------------------------------------------------------------------
Timestamp:        2023-10-14T21:42:18.104Z
Vehicle VIN:      1G1YY2D76P510XXXX
GNSS Fix:         Lat: 45.5152° N, Long: -122.6784° W
Elevation:        18.4 meters
Heading:          184.2° (South)
Vehicle Speed:    41.2 mph (66.3 km/h)
Throttle Input:   14.8%
Hydraulic Line P: 0 PSI
Lateral Accel:    0.04 g
Long Accel:      -0.02 g
Seatbelt Sensor:  Buckled (Driver, Front Passenger)
Odometer:         14,291.8 miles
Active DTC Codes: None
Transmission:     Drive (Gear 4)
----------------------------------------------------------------------

High-frequency location tracking was accompanied by the mechanical harvesting of behavioral signals:

Longitudinal deceleration events (Braking): Monitored through wheel speed pulse counters and hydraulic brake-line pressure transducers. Decelerations exceeding roughly 7.5 miles per hour per second (approximately 0.34g to 0.40g) were flagged automatically in flash memory as "hard braking."
Longitudinal acceleration events: Triggered when the throttle position sensor reported rapid opening or when engine output resulted in positive velocity changes exceeding approximately 7.5 to 8.0 mph per second.
Velocity threshold exceptions: Instantaneous logging of vehicle velocity against hardcoded limits, typically categorizing any travel above 80 miles per hour as an actuarial infraction regardless of the posted highway limit.
Temporal metadata: Absolute timestamps for every ignition-on and ignition-off sequence, exposing travel during late-night windows (typically midnight to 4:00 AM) that underwriters correlate with higher collision claims.

The TCU stored these metrics inside local ring buffers within the vehicle's onboard flash memory. When the car re-entered areas with strong cellular connectivity, the system compiled these discrete data points into JSON or binary payloads and transmitted them via the cellular modem to GM's cloud platform using secure protocols like MQTT and HTTPS.

Drivers had no dashboard lights, infotainment indicators, or instrument cluster displays to signal that this persistent data extraction was taking place. The CAN bus simply kept talking, and the telematics module kept listening.

Actuarial Arbitrage: How LexisNexis and Verisk Turned Sensor Data into Cash

The collection of raw sensor data is technically demanding, but raw CAN bus payloads are useless to an insurance underwriter. An insurance actuary sitting at State Farm, Progressive, Nationwide, or Liberty Mutual cannot feed millions of three-second latitude and longitude timestamps into a risk model.

The raw telematics had to be translated into credit-like risk scores. To build this monetization bridge, GM turned to the two undisputed heavyweights of insurance data warehousing: LexisNexis Risk Solutions and Verisk Analytics.

                 RAW TELEMATICS TO ACTUARIAL SCORE PIPELINE

 +--------------------------------------------------------------------+
 |                    RAW INGESTION (GM TELEMETRICS)                  |
 |  - Hundreds of thousands of trip records per vehicle               |
 |  - 3-second GNSS breadcrumbs                                       |
 |  - Longitudinal/lateral accelerometer vectors                      |
 +---------------------------------+----------------------------------+
                                   |
                                   v
 +--------------------------------------------------------------------+
 |                   NORMALIZATION & FACTOR SCORING                   |
 |  (LexisNexis Telematics Exchange / Verisk DriveInformed)           |
 |                                                                    |
 |  Trip Metrics Analyzed:                                            |
 |  * Hard Brake Rate = (Brake Events / 100 Miles)                    |
 |  * Rapid Acceleration Index = (Accel Events / Total Drive Time)    |
 |  * High-Speed Exposure = (% of Total Distance > 80 mph)            |
 |  * Circadian Risk Weighting = (Miles Driven between 00:00 - 04:00) |
 |  * Geolocation Aggregation = Density of trips in high-loss zones   |
 +---------------------------------+----------------------------------+
                                   |
                                   v
 +--------------------------------------------------------------------+
 |             ACTUARIAL OUTPUT: CONSUMER REPORTING PRODUCT           |
 |  - LexisNexis Telematics Risk Score (Scale: 100 to 999)            |
 |  - FCRA Driving Behavior Report                                    |
 +---------------------------------+----------------------------------+
                                   |
                                   v
 +--------------------------------------------------------------------+
 |                     UNDERWRITING DECISION ENGINE                   |
 |  - Base Rate Multiplier calculated via GLM (Generalized Linear     |
 |    Model)                                                          |
 |  - Unprompted renewal rate increases applied (e.g., +15% to +45%)  |
 |  - Denial of coverage or tier demotion                             |
 +--------------------------------------------------------------------+

LexisNexis Risk Solutions operates a centralized automotive infrastructure known as the Telematics Exchange. According to documents unsealed during the multi-state investigation, the Telematics Exchange was designed to act as an agnostic utility for the insurance industry, ingesting connected-car telemetry from original equipment manufacturers (OEMs) and normalizing the diverse outputs of different vehicle brands into standardized actuarial units.

GM signed enterprise data-licensing agreements with both LexisNexis and Verisk. Under these contracts, GM's cloud servers pushed batch files containing full vehicle histories to the brokers.

The files were anchored by the vehicle's unique 17-character Vehicle Identification Number (VIN). Once LexisNexis ingested a batch, its identity-resolution engines matched the VIN against owner registries, state motor vehicle records, and consumer credit files to bind the telematics directly to an individual driver’s identity.

Verisk Analytics ran a parallel operation under its DriveInformed platform, which at its peak held driving behavior data spanning more than 260 billion miles across eight million vehicles.

Inside these analytics warehouses, algorithms processed the physical realities of driving through generalized linear models (GLMs):

A driver braking sharply to avoid a distracted pedestrian or stopping suddenly because a traffic light cycled unexpectedly fast was stripped of context. The sensor recorded only the negative acceleration vector.
A driver accelerating decisively down an unusually short highway entrance ramp was logged as demonstrating "aggressive vehicle control."
A nurse, factory worker, or hospitality employee driving home from a swing shift at 1:30 AM was penalized through late-night driving penalty coefficients.

The output was a commercial document: the Telematics On-Demand Driving Behavior Report, regulated under the Fair Credit Reporting Act (FCRA) as a consumer investigative report.

When motorists shopped for auto insurance or when their existing carrier initiated an automated renewal pull, underwriting algorithms queried LexisNexis or Verisk. The carrier's pricing engine parsed the score, and if the driver's telematics exceeded internal risk thresholds, the system automatically bumped their rate into higher premium tiers or canceled discounts entirely.

The economic damage to consumers was immediate, compounding, and nearly impossible to trace.

In one high-profile case highlighted in consumer litigation, Kenn Dahl, a software company owner in Washington state, discovered that his insurance premiums jumped by 21 percent. He had maintained an unblemished driving record without an accident or ticket for decades.

After demanding his consumer file from LexisNexis under FCRA disclosure rights, he received an envelope containing a 640-page dossier. The document detailed 640 separate trips taken in his 2021 Chevrolet Bolt EV, complete with precise start times, end times, distances traveled, and dozens of flags for sharp accelerations and hard braking events—all without his explicit knowledge.

In Florida, a driver named Romeo Chicco found himself outright rejected by seven different insurance carriers when seeking coverage for his Cadillac CT5. When he finally secured a policy, the premium was nearly double what he anticipated.

A subsequent inspection of his LexisNexis report revealed that GM had transmitted data on 258 distinct trips he had taken over a six-month period, categorizing standard urban navigation as erratic performance.

While drivers watched their personal monthly expenses escalate, GM was booking pure margin.

State investigators documented that GM received payouts under its data agreements, extracting millions of dollars directly from the brokers. In California alone, Attorney General Rob Bonta's investigators proved that GM had pocketed approximately $20 million from data sales during the four years the scheme ran unchecked.

Dealership Dark Patterns: The Machinery of Manufactured Consent

The central defense mounted by General Motors throughout early regulatory scrutiny was that drivers had agreed to participate. The company pointed to the OnStar user agreement and privacy terms, arguing that customers opted into the Smart Driver feature.

The multi-state investigation spearheaded by Rayfield and Brown exposed that this supposed consumer agreement was manufactured through structural coercion, interface dark patterns, and dealership commission schemes.

                    DEALERSHIP DELIVERY ROOM ENROLLMENT FLOW
                           (The Dark Pattern Funnel)

 [ Dealership F&I Office / Sales Delivery ]
                   |
                   | Sales rep pressured to hit 100% OnStar onboarding target
                   | to unlock monthly corporate bonus pool
                   v
 [ iPad / Touchscreen Interface Presented to Buyer ]
                   |
                   | Pre-checked "Enroll All" Button or Unified Acceptance:
                   | "Click here to accept Emergency Services, Navigation & Safety"
                   v
 [ Bundling Strategy (No Granular Opt-Outs) ]
                   |
                   +---> Real-time Location Sharing (OnStar Crash Response)
                   +---> Remote Key Fob App Permissions (MyChevrolet / MyGMC)
                   +---> OnStar Smart Driver Tracking (Data shared with brokers)
                   |
                   v
 [ Sales Representative Action ]
                   |
                   | In numerous documented cases, the dealer rep tapped 
                   | "Accept All" on behalf of the customer while configuring 
                   | the vehicle's infotainment system during final handover
                   v
 [ Consumer Status ]
                   |
                   +---> Walks out believing they activated roadside assistance
                   +---> Reality: Inadvertently signed blanket telematics waiver
                   +---> Reality: Over 370,000 drivers told "Not Enrolled" via app
                         while TCU actively harvested and transmitted trip logs

For most car buyers, the OnStar enrollment process does not take place in the quiet of their homes after a careful review of digital terms. It happens on the showroom floor or inside the Finance and Insurance (F&I) office at the moment of vehicle delivery.

The customer has just negotiated pricing, reviewed loan terms, signed dozens of legal documents, and is sitting in the driver's seat waiting for the sales representative to demonstrate the dashboard features.

Investigators discovered that General Motors established high-stakes financial incentive programs for its dealer network that tied dealership profitability to OnStar enrollment metrics. Dealerships and sales associates were paid direct bonuses for hitting high percentages of OnStar activations during the vehicle handover phase.

If a dealership allowed new vehicle buyers to drive off the lot without activating their OnStar accounts, the dealership's metrics dropped, reducing manufacturer allocations of popular, high-margin vehicle inventory.

This corporate quota system created an aggressive environment where proper consent was an obstacle to overcome. Dealership personnel utilized tablets and showroom workstations running a single, bundled sign-up workflow.

Instead of presenting the Smart Driver program as an optional, independent telematics tracking program with a distinct privacy disclosure, the software architecture bundled the program alongside fundamental safety and convenience services:

Automatic Crash Response (which alerts 911 dispatchers if an airbag deploys).
Stolen Vehicle Assistance (which allows law enforcement to track and disable a stolen vehicle).
Turn-by-Turn Navigation and Roadside Assistance.
Remote Key Fob features embedded in the MyChevrolet, MyGMC, MyBuick, and MyCadillac smartphone apps.

Sales representatives routinely utilized "dark patterns"—user interface designs intentionally crafted to manipulate users into making choices against their own interests.

The dealership software presented unified "Accept All" prompts where refusing Smart Driver required navigating multi-step sub-menus that delayed delivery. In widespread instances documented by state investigators, dealership employees clicked through the terms and conditions on their tablets without showing the screens to the consumers at all, completing the enrollment on the customer’s behalf.

The multi-state investigation uncovered an even more serious administrative failure: a technical disconnect that rendered consumer status entirely opaque.

For more than 370,000 drivers, the OnStar and brand mobile applications explicitly stated that they were "Not Enrolled" in the Smart Driver program. When motorists checked their profile settings in the app, the toggle button was visually switched off.

Yet inside the vehicle's TCU, the telemetry capture task remained active. GM’s data pipeline continued to log their drive histories, package their locations, and transfer the records to LexisNexis and Verisk.

Internal documents obtained by investigators showed that when GM engineers and product managers eventually discovered this discrepancy, the company opted not to notify affected customers. There was no system-wide alert, no push notification, and no email admitting the operational error.

Instead, GM quietly altered the back-end database entries to unenroll those 370,000 drivers, leaving years of previously harvested trip records in the commercial databases of the data brokers and insurance companies, where they continued to impact renewal rates.

When consumers tried to unenroll intentionally, the process was intentionally cumbersome. While activating the service was a frictionless, one-touch action performed at the dealership, opting out required navigating a complex seven-step process buried inside smartphone app submenus or placing a phone call to OnStar customer service representatives trained to talk drivers out of canceling.

Mary Barra’s Software Gamble: The Corporate Strategy Behind the Data Harvest

To understand how GM found itself facing a $220 million reckoning across 36 states, one must examine the corporate narrative Detroit sold to Wall Street between 2018 and 2022.

For more than a century, automotive manufacturing has been a low-margin, capital-intensive, cyclical business. Automakers typically trade at price-to-earnings (P/E) multiples of 5x to 8x.

By contrast, Silicon Valley technology firms and Software-as-a-Service (SaaS) providers trade at multiples of 25x to 40x because their revenues are recurring, subscription-based, and unburdened by steel stamping, tooling costs, and union labor.

Under Chief Executive Officer Mary Barra, General Motors set out to convince institutional investors that it was not an industrial metal-bender, but a next-generation software platform.

 +-------------------------------------------------------------------+
 |            DETROIT'S SOFTWARE VALUATION TRANSFORMATION TARGET     |
 +-------------------------------------------------------------------+
 |  Traditional Automaker Model         | Tech / SaaS Platform Model |
 |  - P/E Multiple: 5x – 8x             | - P/E Multiple: 25x – 40x  |
 |  - One-time transaction at sale      | - Continuous recurring ARR |
 |  - Low gross margins (8% – 12%)      | - High software margins    |
 |  - Cyclical inventory risk           |   (70% – 85%)              |
 +--------------------------------------+----------------------------+
                                        |
                 GM 2021 INVESTOR DAY PLEDGE:
                 "Unlock $20B to $25B in annual recurring 
                  software and services revenue by 2030"
                                        |
                                        v
 [ STRATEGIC PILLAR: CONNECTED VEHICLE DATA PLATFORM ]
 * Monitored via GM Fleet and OnStar
 * Direct monetization via Data Broker Contracts
 * Commercialization of GM driver location data
 * Behavioral Telematics sold to Actuarial Exchanges

The peak of this corporate repositioning arrived during GM’s Investor Day in October 2021.

Barra and her executive leadership team outlined an ambitious financial target: GM pledged to double its annual corporate revenues to approximately $280 billion by 2030, with $20 billion to $25 billion of that total coming directly from software and connected vehicle services.

The company presented investors with visions of high-margin recurring income generated by in-vehicle subscriptions, infotainment software upgrades, Super Cruise hands-free driving licenses, and corporate data commercialization.

Data was celebrated as the new oil of the automotive enterprise. Inside the Renaissance Center in Detroit, cross-functional teams were assembled within GM’s digital business divisions to identify untapped data streams that could be monetized immediately, long before fully autonomous vehicles reached commercial maturity.

The CAN bus was generating hundreds of gigabytes of raw behavioral data per fleet vehicle every year. Connected vehicle divisions recognized that the insurance industry was desperate for real-world driving data.

For decades, insurance companies had relied on crude, static proxies to price automotive risk:

The driver's age and marital status.
The zip code where the car was garaged.
The driver's credit score (which underpins insurance risk tiers in most states).
Prior accidents and moving violations reported through the Department of Motor Vehicles.

These static proxies were imperfect. Insurers understood that dynamic risk—actual throttle use, braking intensity, adherence to speed limits, and night-time mileage—provided predictive power that blew historical actuarial tables apart.

Insurers had tried tracking drivers via OBD-II dongles (such as Progressive's Snapshot or State Farm's Drive Safe & Save) or through smartphone applications. But dongles were expensive to mail out, and consumers routinely unplugged them when they intended to drive fast.

Smartphone apps were plagued by battery drain and had difficulty distinguishing whether the phone's owner was driving, riding as a passenger in an Uber, or sitting on a commuter bus.

The connected car solved every technical hurdle the insurance industry faced.

The hardware was permanently integrated into the vehicle chassis.
The sensors were calibrated to the machine’s physical weight and braking dynamics.
The power supply was drawn from the car's alternator and 12-volt battery.
The driver could not disconnect the module without voiding warranties or bricking their infotainment screen.

By positioning itself as an institutional pipeline of behavioral telematics, GM could establish long-term recurring revenue relationships with risk aggregators. The $20 million that GM pulled from LexisNexis and Verisk was the vanguard of a broader data economy.

The pressure from top corporate brass to show immediate digital revenue velocity encouraged product managers to prioritize data ingestion over user transparency, pushing the legal limits of consumer consent until the technical pipeline collided with state consumer protection statutes.

The Legal and Regulatory Unraveling

The architecture began to collapse in March 2024, when investigative reporter Kashmir Hill published an expose in The New York Times documenting the direct link between GM’s OnStar Smart Driver system and unexplained insurance rate increases experienced by vehicle owners.

The fallout was immediate and catastrophic for GM’s public relations and legal departments. Within days, thousands of motorists downloaded their consumer disclosure files from LexisNexis and Verisk, discovering that their day-to-day driving habits had been tracked, logged, and quantified.

                          TIMELINE OF REGULATORY & LEGAL ACTION
                                   (2024 – 2026)

  March 2024:
  The New York Times breaks investigation exposing GM's data pipeline to 
  LexisNexis and Verisk. Motorists discover comprehensive driving dossiers.
        |
  March 20, 2024:
  GM publicly announces it has severed data-sharing partnerships with 
  LexisNexis Risk Solutions and Verisk Analytics.
        |
  April – June 2024:
  Consumers file dozens of federal class actions across the country; GM 
  announces the permanent discontinuation of OnStar Smart Driver.
        |
  August 2024:
  Texas Attorney General Ken Paxton files an enforcement lawsuit against GM 
  alleging deceptive trade practices affecting 1.5 million Texas drivers.
        |
  January 2025 – January 2026:
  The FTC finalizes a 20-year consent order against GM, banning it for 
  five years from selling consumer telematics to credit reporting agencies.
        |
  April 2026:
  U.S. District Judge Thomas Thrash upholds federal wiretap and FCRA claims 
  in nationwide multidistrict litigation (In re: Consumer Vehicle Driving 
  Data Tracking Litigation).
        |
  June 2026:
  California AG Rob Bonta fines GM $12.75 million for violating the 
  California Consumer Privacy Act (CCPA).
        |
  October 29, 2026:
  Bipartisan coalition of 36 state Attorneys General, co-led by Oregon and 
  Washington, announces the landmark $220 million multistate settlement.

On March 20, 2024, GM announced that it had abruptly severed its data-sharing partnerships with both LexisNexis and Verisk. By June, the automaker announced the complete discontinuation of the Smart Driver program across all brand divisions.

Halting the data pipeline did not extinguish the regulatory scrutiny. Instead, it provided prosecutors and state attorneys general with a clear evidentiary baseline.

The legal confrontation unfolded across four major fronts:

1. The FTC Investigation and Federal Consent Decree

In January 2025, the Federal Trade Commission, led by Chair Lina Khan, initiated enforcement proceedings against General Motors LLC, General Motors Holdings, and OnStar LLC.

The FTC’s formal complaint established that GM had engaged in unfair and deceptive business practices under Section 5(a) of the FTC Act. The FTC documented the three-second geolocation tracking intervals and detailed how GM failed to maintain meaningful boundaries over the downstream exploitation of that data.

The FTC finalized a 20-year consent order that imposed a five-year ban prohibiting GM from disclosing any consumer geolocation or driver behavior data to consumer reporting agencies.

Crucially, the FTC order established new national standards for automotive telemetry, mandating that GM must provide an easily accessible mechanism for consumers to request the total deletion of previously harvested data and guarantee that drivers can disable geolocation tracking entirely without losing fundamental vehicle functionality.

2. State Enforcement: Texas, California, and Beyond

Individual state attorneys general moved rapidly under their sovereign consumer protection authorities.

In August 2024, Texas Attorney General Ken Paxton filed a massive lawsuit alleging that GM deceived more than 1.5 million Texans through predatory data collection disguised as vehicular safety tools.

In June 2026, California Attorney General Rob Bonta slapped GM with a $12.75 million civil penalty—the largest financial sanction ever handed down in the history of the California Consumer Privacy Act (CCPA).

Bonta’s office focused on the principle of "purpose limitation," proving that collecting telematics for crash notification and then repurposing those same data streams for actuarial monetization violated California law.

3. The Federal Multidistrict Class Action (MDL 3115)

Simultaneously, dozens of private class-action lawsuits filed by vehicle owners were consolidated by the Judicial Panel on Multidistrict Litigation into a single master proceeding: In re: Consumer Vehicle Driving Data Tracking Litigation, No. 1:24-md-03115, assigned to U.S. District Judge Thomas Thrash in the Northern District of Georgia.

GM and the data brokers filed motions to dismiss, arguing that the claims were blocked by mandatory arbitration clauses embedded in OnStar terms and conditions, and asserting that auto telematics do not constitute "wire communications" protected by federal surveillance statutes.

In a 200-page decision in April 2026, Judge Thrash rejected GM's core defense.

The court ruled that the plaintiffs had plausibly alleged claims under the Federal Wiretap Act and the Stored Communications Act, reasoning that the secret, real-time routing of in-vehicle communications to third-party corporate entities could constitute an unlawful interception under federal law.

The ruling allowed claims for civil conspiracy, unjust enrichment, and Fair Credit Reporting Act violations to proceed directly toward a jury trial.

4. The 36-State Bipartisan Coalition

While individual states litigated and private plaintiffs pursued discovery, an executive committee of state attorneys general began negotiating a nationwide settlement.

Co-led by Oregon and Washington, and supported by an executive group that included Colorado, Connecticut, Delaware, Illinois, Kentucky, Maryland, Minnesota, New Jersey, Utah, and Vermont, the coalition pooled investigative resources.

The state AGs possessed a distinct advantage: unlike federal regulators limited by specific statutory remits, state attorneys general wielded broad Unfair and Deceptive Acts and Practices (UDAP) statutes.

These laws allow states to punish deceptive commercial behavior that occurred well before modern consumer privacy statutes were enacted. This culminated in the $220 million settlement finalized on October 29, 2026.

Terms of the $220 Million Multistate Resolution

The $220 million settlement represents a seismic structural disruption to the automotive business model.

The financial penalty is distributed across the participating states based on population and connected-vehicle sales density. Washington state will receive $12.3 million from the recovery, Oregon will receive more than $10.4 million, and California, Texas, Ohio, and New York will secure tens of millions more to fund ongoing state consumer protection and privacy enforcement divisions.

The operational, technical, and structural mandates imposed upon GM under the settlement terms go far beyond the monetary penalty:

Mandatory Affirmative Express Consent

The days of buried contractual disclosures are over.

The agreement strictly forbids GM from collecting, uploading, or sharing consumer driving behavior or location metrics without first securing affirmative, unbundled, express consent.

Under the decree, consent cannot be bundled into general vehicle sales contracts, finance agreements, warranty enrollments, or overarching OnStar terms. If GM wishes to harvest telematics, it must present a standalone, distinct disclosure that clearly explains what data is harvested, how often it is polled, and precisely which third-party commercial entities will receive it.

                 PERMISSIBLE VS. PROHIBITED ENROLLMENT MODELS
                       UNDER MULTISTATE SETTLEMENT

  PROHIBITED (2016 - 2024):
  +---------------------------------------------------------------+
  | [X] I agree to OnStar Safety, Navigation, Vehicle Diagnostics |
  |     and Smart Driver Programs.                                |
  |                                                               |
  | (Single checkbox hidden within 40-page purchase documentation)|
  +---------------------------------------------------------------+
                                 VS.
  MANDATORY (POST-SETTLEMENT):
  +---------------------------------------------------------------+
  | CRASH & EMERGENCY SAFETY SERVICES                             |
  | [ ACCEPT ]   [ DECLINE ]                                      |
  |                                                               |
  | ------------------------------------------------------------- |
  | BEHAVIORAL TELEMATICS & VEHICLE SENSOR COLLECTION             |
  | Samples speed, braking, acceleration, and precise GPS every   |
  | three seconds. This data will be shared with outside parties. |
  | [ ACCEPT ]   [ DECLINE ]                                      |
  |                                                               |
  | *Declining this feature CANNOT alter your vehicle warranty,   |
  |  disable physical safety systems, or impede navigation.       |
  +---------------------------------------------------------------+

Decoupling of Hardware and Core Vehicle Functions

The settlement codifies an engineering safeguard: GM is barred from conditioning physical vehicle operations on data tracking.

A driver who chooses to decline all tracking must retain complete, unhindered access to the vehicle's engine performance, physical braking controls, comfort systems, safety systems, infotainment, and warranty coverage.

GM cannot implement digital throttling, display recurring nag-screens, or degrade the performance of the automobile as punishment for preserving privacy.

Dealership Compliance Auditing

Recognizing that the showroom floor was the operational epicenter of fraudulent enrollments, the settlement binds GM to strict dealership oversight protocols.

GM must redesign all showroom point-of-sale software to prevent dealership employees from accepting privacy policies on behalf of consumers.

The automaker must deploy independent mystery-shopper audit programs to verify that dealership sales and finance personnel are not engaging in dark patterns, misrepresenting the nature of OnStar, or pre-activating telematics profiles during the pre-delivery inspection phase.

Compulsory Purge and Verification Protocols

GM must construct a permanent, user-accessible consumer portal that allows any vehicle owner to view what driving telemetry GM currently retains on their VIN or identity.

Consumers must be granted a permanent "Right to Purge" mechanism that compels GM to delete historical vehicle location and driving records from its servers within 30 days of a verified request.

Furthermore, GM must certify that it has notified all third-party commercial data brokers that ingested this information between 2016 and 2024 that the data was obtained without verified affirmative consent.

Independent Algorithmic and Privacy Audits

For the next decade, GM must submit its internal telematics architecture and cloud software deployments to regular privacy compliance reviews conducted by independent third-party assessors.

These audit reports must be delivered directly to the executive committee of state attorneys general, creating an ongoing layer of technical accountability.

Technical Vectors: The Data Points That Built Your Risk Score

To understand how GM driver location data and trip telematics were weaponized against vehicle owners, it is essential to examine the specific technical vectors captured by the TCU.

During the multi-state investigation, forensic examiners cataloged the primary data fields extracted by the OnStar Smart Driver system and translated by data brokers into risk scores:

+-----------------------------------------------------------------------------------------------------------------------+
| SENSOR INPUT                  POLLING FREQUENCY  CAN BUS SOURCE               ACTUARIAL EXPOSURE                      |
+-----------------------------------------------------------------------------------------------------------------------+
| Geolocation (Lat/Long)        Every 3 seconds    High-Precision GNSS Receiver Commute routing, home/work locations,   |
|                                                                               trips to high-theft or high-loss zones  |
|                                                                                                                       |
| Longitudinal Deceleration     Event-driven       Hydraulic Brake Line Transducer / Heavy penalty for >0.34g stops;    |
| (Hard Braking)                (Continuous poll)  Wheel Speed Pulse Counters   treated as tailgating/inattention       |
|                                                                                                                       |
| Longitudinal Acceleration     Event-driven       Throttle Position Sensor /   Penalized as aggressive control or      |
| (Rapid Acceleration)          (Continuous poll)  Transmission Output Shaft    unnecessary risk-taking                 |
|                                                                                                                       |
| Velocity Exceptions           Continuous         ABS / Wheel Speed Pulse      Immediate risk tier degradation if      |
|                               (Calculated)       Counters                     travel exceeds 80 mph                   |
|                                                                                                                       |
| Circadian Risk Factor         Ignition-On to     Engine ECU Internal Clock /  Heavily weighted penalty for driving    |
| (Night Driving)               Ignition-Off       GPS Universal Time Fix       between 12:00 AM and 4:00 AM            |
|                                                                                                                       |
| Dynamic Seatbelt State        Continuous         Airbag Diagnostic Sensing &  Direct risk flag indicating failure to  |
|                               (State change)     Diagnostic Module (SDM)      adhere to baseline safety protocols     |
+-----------------------------------------------------------------------------------------------------------------------+

When packaged into an FCRA Driving Behavior Report, these discrete sensor variables were aggregated into composite risk indices.

Insurers ingested these metrics to generate a risk multiplier. A vehicle owner with an otherwise clean driving record who logged 1,000 miles a month could see their base insurance premium multiplied by a factor of 1.25x or 1.40x if the car registered an average of more than three "hard braking" events or two "rapid acceleration" events per 100 miles.

The algorithms ignored real-world context:

A driver avoiding an out-of-control vehicle by accelerating hard was penalized.
A parent hitting the brakes firmly because a child stepped into a suburban crosswalk was logged as a negative driver.
A worker operating a night shift was structurally penalized simply for being on the road during hours insurers correlate with higher collision claims.

The continuous extraction of GM driver location data also exposed granular private travel histories.

Because the system polled GPS coordinates every three seconds, an analyst with access to the raw database could reconstruct exact driving routes.

This breadcrumb trail revealed sensitive lifestyle details: visits to specialized medical clinics, oncologists, addiction rehabilitation centers, religious facilities, union halls, and romantic partners.

All of this was packaged without consumer awareness, bundled under the guise of an internal program called "Smart Driver" that claimed to offer safety tips.

The Broader Auto Industry Tremor

General Motors is not an isolated actor in connected-vehicle surveillance; it was simply the first manufacturer caught red-handed with an established data monetization pipeline.

The $220 million settlement sends shockwaves through every automotive executive suite in North America, Europe, and Asia.

In late 2023, the Mozilla Foundation released a comprehensive privacy study evaluating 25 major car brands, concluding that modern automobiles were officially "the worst product category for privacy" Mozilla had ever reviewed.

The report established that every major automaker surveyed collected excessive amounts of personal data—ranging from location metrics and cabin microphone recordings to behavioral telematics—with 84 percent of brands retaining the contractual right to share or sell customer data to third-party brokers, service providers, and law enforcement.

                  INDUSTRY-WIDE CONNECTED VEHICLE RISK PROFILE
               (Automotive Telematics Operations Under Scrutiny)

  +-------------------+----------------------------------------------------+
  | AUTOMAKER         | TELEMATICS PLATFORMS / INVESTIGATIVE SCRUTINY      |
  +-------------------+----------------------------------------------------+
  | General Motors    | OnStar Smart Driver (Settled for $220M;            |
  |                   | discontinued partnerships with LexisNexis/Verisk)  |
  |                   |                                                    |
  | Ford Motor Co.    | FordPass Connect / Commercial Telematics;          |
  |                   | Facing scrutiny over insurer data-sharing terms    |
  |                   |                                                    |
  | Stellantis        | Uconnect Box / Ram, Jeep, Dodge Telematics;        |
  |                   | Auditing data exchange broker integration points   |
  |                   |                                                    |
  | Honda & Acura     | HondaLink; Subject to state regulatory inquiries   |
  |                   | regarding DriverFeedback score monetization        |
  |                   |                                                    |
  | Hyundai & Kia     | Bluelink / Kia Connect Driving Score systems;      |
  |                   | Scrutinized for dealer-floor onboarding workflows  |
  |                   |                                                    |
  | Toyota & Lexus    | Toyota Connected Services / Drive Pulse;           |
  |                   | Currently restructuring opt-in consent pathways    |
  +-------------------+----------------------------------------------------+

Virtually every major manufacturer operating in North America has experimented with connected-car data programs:

Ford Motor Company operates FordPass Connect, offering telematics integration programs that have attracted regulatory scrutiny regarding data transfers to third-party commercial aggregators.
Honda and Acura have marketed DriverFeedback programs through the HondaLink ecosystem, utilizing smartphone and onboard sensors to score driving behavior.
Hyundai and Kia have integrated Driving Score algorithms into their Bluelink and Kia Connect platforms, facing similar questions regarding dealership onboarding practices.
Toyota and Lexus manage Toyota Connected Services and the Drive Pulse platform, analyzing trip events and assigning driving performance ratings.
Stellantis (encompassing Ram, Jeep, Dodge, and Chrysler) has built Uconnect systems capable of uploading live telemetry back to cloud clearinghouses.

In the wake of GM’s legal defeat, legal counsels at competing automakers are executing emergency compliance audits.

Automakers are reviewing data licensing contracts, scrubbing third-party Software Development Kits (SDKs) from infotainment units, and overhauling connected-services enrollment terms.

Data brokers are facing parallel fallout. Verisk Analytics shut down its automaker-sourced driving data clearinghouse entirely in mid-2024 as the litigation intensified.

LexisNexis Risk Solutions has faced a wave of individual and class-action lawsuits challenging the legality of its Telematics Exchange under the Fair Credit Reporting Act and federal wiretapping statutes.

State attorneys general have established an enforcement precedent: where the federal legislative branch has failed to pass comprehensive federal privacy laws, state AGs will step into the regulatory void, using their state consumer protection powers to police connected vehicle tracking.

Automakers can no longer rely on sixty-page online privacy policies to immunize their surveillance practices.

What Comes Next: The Post-Telematics Reality

The resolution of the multistate investigation marks the end of an era for unrestrained automotive data monetization, but it leaves critical technical, financial, and legal questions unresolved.

The primary battleground now shifts to federal court in Atlanta.

The $220 million multistate settlement resolves state civil penalty actions, consumer protection investigations, and regulatory enforcement claims. It does not extinguish the claims of private consumers seeking restitution in In re: Consumer Vehicle Driving Data Tracking Litigation.

Because Judge Thrash upheld federal wiretap, stored communications, and FCRA causes of action, GM, LexisNexis, and Verisk face the prospect of a nationwide class trial.

Plaintiffs' attorneys are seeking billions of dollars in statutory damages, demanding direct financial compensation for millions of drivers whose insurance premiums climbed because their vehicle data was commercialized without valid consent.

                THE FUTURE OF IN-VEHICLE DATA ARCHITECTURE
       (Shift from Centralized Harvesting to Edge-Compute Privacy)

  OLD PARADIGM (2016 – 2024):
  Vehicle CAN Bus ---> Telematics Unit ---> GM Cloud ---> Data Brokers ---> Insurers
  * Raw sensor extraction (Speed, Braking, 3-sec GPS coordinates)
  * Centralized harvesting and third-party monetization
  * Downstream underwriting penalties applied without driver knowledge

  NEW PARADIGM (2026 and Beyond):
  Vehicle CAN Bus ---> In-Cabin Edge Processing (Local ML Model)
                             |
                             v
  +---------------------------------------------------------------+
  | On-Device Edge Compute (Local Enclave)                        |
  | - Telematics data processed locally inside the vehicle        |
  | - Raw GM driver location data NEVER leaves the local machine  |
  | - Privacy Sandbox / Zero-Knowledge Proofs for verification   |
  +---------------------------------------------------------------+
                             |
      Driver Grants Explicit, Verifiable, Unbundled Opt-In
                             v
  Verified Abstracted Token: "Tier 1 Safe Driver Certified"
  (No GPS logs, no trip timestamps, no raw braking vectors transmitted)

From an engineering perspective, Detroit's approach to the software-defined vehicle must undergo a fundamental redesign.

Automakers cannot abandon connected services entirely; consumers demand navigation, over-the-air operating system updates, remote battery management for electric vehicles, and instant crash assistance.

To deliver these digital conveniences without violating the new regulatory mandates, automotive systems engineers are pivoting toward edge computing and zero-knowledge data architectures:

Instead of uploading raw CAN bus telemetry and three-second GPS breadcrumbs to the cloud, modern TCUs are being engineered to process metrics locally inside secure hardware enclaves.
If a driver elects to participate in a usage-based insurance (UBI) discount program, the vehicle's onboard computer calculates a verified risk token locally on the device.
The car never transmits raw GPS histories, trip itineraries, or mechanical brake-pressure timestamps to the manufacturer or an external broker.
It transmits only an abstracted cryptographic confirmation certifying that the vehicle operated within agreed safety parameters, keeping raw GM driver location data isolated within the vehicle's onboard memory.

The economic consequences for Detroit will reverberate for years.

The software-driven multi-billion-dollar recurring revenue targets announced by automakers between 2020 and 2022 were modeled on frictionless, unconstrained data commercialization.

Now that thirty-six state attorneys general have erected strict regulatory, technical, and operational barricades, those software financial projections will have to be revised downward.

Consumers have delivered an unmistakable message through state regulators: they are willing to buy technologically advanced, connected automobiles, but they will not accept their vehicles operating as mobile tracking infrastructure for the insurance industry.

The $220 million settlement serves as a binding precedent.

Any automaker that treats a private car as a vehicle for non-consensual surveillance risks facing the full weight of multi-state regulatory enforcement, structural business disruption, and catastrophic brand erosion.

As connected cars evolve into rolling supercomputers, the boundaries drawn by this landmark resolution ensure that the person holding the steering wheel remains the driver—not the product.