Car-insurance telematics / surveillance pricing — UBI apps, the covert automaker→LexisNexis/Verisk→insurer data pipeline, FTC v. GM, and Arity's phone-SDK tracking (both sides, dated)

Built 2026-06-28 from research/spec-telematics-insurance.json. The auto/surveillance/insurance nexus (sub-block of the auto epic #177). Cross-links GM, FTC, State_AGs, US_Auto_Makers, and the surveillance/ownership-erosion thread.

Frame. Usage-based insurance (UBI) telematics (Progressive Snapshot, Allstate Drivewise, State Farm, Root) tracks speed/braking/miles/phone-use/location via dongles or always-on phone apps. The bigger story is the covert pipeline: automakers (GM OnStar "Smart Driver"; also Honda/Hyundai/Kia) sold driving + geolocation data to data brokers LexisNexis Risk Solutions + Verisk, who repackaged it into insurance risk scores → premium hikes without consent (NYT, Mar 2024). Enforcement followed: FTC v. GM/OnStar (order finalized Jan 2026), California AG ($12.75M), Texas AG v. Allstate/Arity (Jan 2025). Discipline. Products, the covert sales, and the enforcement are fact. The two supplied framings ("location-inference price discrimination" and "no-alternative collective coercion") are graded — neither adopted as asserted fact nor dismissed. Overlay; excluded from the proofs.

1. The UBI apps (fact)

Insurers offer "discounts" for installing a dongle or always-on phone app that scores driving — Progressive Snapshot, Allstate Drivewise, State Farm Drive Safe & Save, Root. The apps read phone sensors and can run continuously (location/movement even when not driving the insured car). Surveys show most non-users decline over privacy and would only opt in if insurers promised not to sell the data + gave deletion control. Fact (the products + the continuous-collection capability).

2. The covert pipeline (fact)

The scandal (NYT / Kashmir Hill, Mar 2024): automakers collected detailed driving + precise geolocation (GM's OnStar "Smart Driver" as often as every ~3 seconds) and sold it to brokers LexisNexis Risk Solutions and Verisk Analytics, which packaged it into risk reports sold to insurers — and premiums rose on data drivers never knowingly shared (one driver's premium reportedly jumped ~80% from 603 logged trips). Honda, Hyundai, Kia were also implicated. GM made ~$20M and gave consumers no notice. Fact.

3. The enforcement (fact)

1. FTC v. GM + OnStar — order finalized Jan 2026: a 5-year ban on sharing geolocation/driver-behavior data with consumer-reporting agencies + ~20-year consent/transparency requirements.
2. California AG — a $12.75M settlement (CCPA data-minimization/purpose-limitation).
3. Texas AG — sued Allstate + its data subsidiary Arity (Jan 2025) for unlawfully collecting/selling location + movement data from 45M+ trips; Arity embedded SDKs in third-party phone apps (e.g. Life360, GasBuddy) to harvest movement data from people with no insurance relationship.

Fact.

4. The price-discrimination risk (graded)

The first supplied framing: continuous geolocation enables inferring behavior/health and pricing on it (e.g., regular pharmacy visits could proxy a health condition). This is a documented capability/risk the pipeline creates — not an asserted current insurer practice. Per read-words-for-intent: insurers say UBI is voluntary/discount-only and that they "don't sell" data, yet the covert automaker sales + Arity's app-SDK harvesting show data flowing without consent into risk scoring — so stated intent and documented conduct diverge, which is the warranted finding; the specific "punish you for being sick via pharmacy visits" mechanism is plausible-but-unproven. Grade: the price-discrimination capability is documented; a specific health-punishment practice is unverified (flagged, not adopted).

5. The "no alternative" claim (graded)

The second framing — "insurers force telematics by offering no alternative" — is partly true / partly overstated. Non-telematics insurers still exist (so "all force it" is unsupported), but three structural facts make opting out leaky: (a) the shared data-broker layer — LexisNexis/Verisk score drivers across insurers, so declining one company's app doesn't escape the broker score; (b) the covert car pipeline meant your vehicle reported you regardless of any insurance app; (c) the industry trend is telematics-as-default + surveillance-priced books. So the "collective" effect is real via shared infrastructure + parallel conduct — not necessarily coordinated collusion (composition guard: parallel conduct + a shared broker layer ≠ a single "insurer mind"). Grade: "no alternative at all" unsupported; "shared-infrastructure coercion/leakiness" documented.

6. The honest reading

Car-insurance telematics is real surveillance pricing, and the worst of it was covert: automakers (led by GM/OnStar) sold precise driving + geolocation data to LexisNexis/Verisk, who scored drivers for insurers who raised premiums on data consumers never knowingly shared — until the FTC (Jan 2026), California, and Texas (v. Allstate/Arity) intervened. The two concerns grade out as: (1) location-based price discrimination is a documented capability the pipeline enables (a specific health-punishment practice is plausible-but-unproven, flagged not asserted), and (2) the "no alternative" coercion is real not as a literal universal mandate but via the shared data-broker layer + the covert car pipeline (parallel conduct + shared infrastructure, not proven collusion). Read by read-words-for-intent: insurers' "voluntary / we don't sell" framing diverges from the documented covert sales. Both sides retained with dates. Overlay; excluded from the proofs.

Sources: NYT — automakers quietly shared drivers' data with insurers (Mar 2024); FTC — action against GM/OnStar over geolocation/driving data; TechCrunch — FTC GM data order settled (Jan 2026); Hunton — California AG $12.75M GM settlement (CCPA); Texas AG — sues Allstate & Arity (Jan 2025); Edmunds — phone apps selling driving data (Arity SDK).

← Research index · structured data: spec-telematics-insurance.json · spec-telematics-insurance.md