Urban Autonomy Lab
Chapter 4 · EV48

Legitimacy Crisis: Public Trust Data & Hidden Human Operations

Date collected: 26 May 2026
Summary

Surveys show majority distrust AV safety claims; covert remote operators and "safety drivers" mask true autonomy levels, eroding public legitimacy

EV48 — Legitimacy Crisis: Public Trust Data & Hidden Human Operations

Evidence Note for CAM Governance White Paper

Compiled April 2026

Overview

The legitimacy failure mode for CAM is not abstract — it is quantifiable and current. With only 13% of US drivers trusting self-driving vehicles, Waymo operating at 500,000 rides/week in 10 cities, and the Markey report revealing that “autonomous” vehicles are underwritten by unregulated remote human operators, the conditions for a legitimacy crisis are in place. This evidence note consolidates the trust data, identifies the structural drivers of legitimacy deficit, and draws governance implications for the paper’s Legitimacy failure mode and Ch.7 toolkit sections.

Public Trust: Headline Data

  • 13% of US drivers trust self-driving vehicles (WEF / industry surveys, Q1 2026)
  • 61% of US drivers are afraid to ride in a self-driving car
  • Trust has NOT improved materially despite increasing deployment (Waymo has been operating since 2018 in Phoenix; public trust nationally remains low)
  • The gap between actual safety performance and public perception is significant — AVs are likely statistically safer than human drivers in controlled conditions, yet 87% do not trust them

Why trust matters for governance

“Regulation and public acceptance are conditions for deployment, not secondary concerns — operators lack permission more than they lack technology.” (Progressive Policy Institute, 2026)

A 13% trust rate means 87% of potential users are not willing passengers. At current Waymo scale (500K rides/week from a population of 330M), the addressable market is constrained not by supply but by demand — specifically, by legitimacy. Cities that authorise CAM without first addressing the legitimacy gap are authorising a technology that most of their residents will refuse to use.

Sources

Structural Drivers of Legitimacy Deficit

1. The RAO Revelation (see EV45)

The Markey Senate report (March 31, 2026) exposed that marketed “Level 4 autonomous” vehicles rely on remote human operators who can intervene without passenger knowledge. Key legitimacy implications:

  • Passengers booking “autonomous” rides do not know they may be assisted by a human operator in another country
  • Some overseas RAOs do not hold US driver’s licences
  • No company (except Waymo, partially) discloses RAO intervention rates
  • No regulatory requirement for passenger notification of RAO involvement

Legitimacy gap: The marketed product (autonomous travel) differs from the delivered product (human-assisted remote operation). This is a category of legitimacy failure — not safety failure, but transparency failure.

2. Incident Accumulation Without Accountability

California DMV data as of April 2026:

  • 970 AV collision reports received
  • 5,202 total AV incidents reported through November 2025
  • 451 injuries/fatalities across all operators
  • 1,429 Waymo incidents; 117 injuries, 2 fatalities

These numbers exist in a public database, but there is no public-facing dashboard translating them into meaningful safety comparisons. The absence of contextualised safety data — compared to human driver rates per mile, for example — prevents informed public assessment.

Legitimacy gap: Incidents accumulate visibly; safety performance in context is invisible.

3. High-Profile Incidents (Recent)

Austin ambulance blockage (March 2026):

  • Waymo vehicle blocked ambulance response in downtown Austin during active incident
  • No material delay to emergency response; but significant public visibility and media coverage
  • Reinforced perception that AVs are obstacles to emergency services

Wuhan fleet failure (April 1, 2026) (see EV44):

  • 100+ Baidu robotaxis simultaneously stalled on elevated highways; multiple rear-end collisions; passengers trapped
  • Even if technically no fatalities, the visual and media impact was severe
  • Directly reinforces public fear of what happens when autonomous systems fail at scale

Significance for legitimacy: Neither incident caused fatalities, but both confirmed the specific fears that drive the 61% afraid-to-ride figure — that AVs will fail in ways that trap passengers, block emergency services, or cause cascading accidents they cannot prevent.

4. Missing Transparency Infrastructure

Progressive Policy Institute (2026) proposes a transparency framework for restoring public trust:

  • Public-facing safety dashboards: Real-time or periodic publication of safety metrics by operator, by city, per million miles
  • Crash rate comparisons: AV vs. human driver fatality and injury rates per billion vehicle miles travelled — contextualising absolute numbers
  • Researcher data access: Independent academic and safety researcher access to AV operational data
  • Natural language vehicle explanations: Vehicles explaining their decisions in plain language to passengers and bystanders
  • RAO disclosure: Mandatory passenger notification if journey involved remote operator intervention

Governance Implications for the CAM Paper

For Ch.5 — Legitimacy Failure Mode

The legitimacy failure mode is not a future risk — it is present and quantified. 87% of the US adult population does not trust self-driving vehicles. Waymo is operating at scale in that environment. The failure mode activates when a legitimacy deficit — whether triggered by an incident (Cruise 2023, Xiaomi 2025, Wuhan 2026), a transparency revelation (RAO, Markey report), or accumulated distrust — produces regulatory suspension, political opposition, or market rejection that rolls back CAM deployment.

The trust gap also has equity dimensions not yet quantified: survey data on trust by income, race, and geography would reveal whether AVs face particular legitimacy deficits among the communities they most need to serve (first/last mile, transit-adjacent deployment, underserved areas).

For Ch.7 — Toolkit: Rules & Instruments

Transparency instruments for the legitimacy toolkit:

  1. Mandatory public safety dashboards: City operating licences should require quarterly publication of safety data (incidents per million miles, injury rates, emergency response interaction rates) in standardised, publicly accessible format
  2. Contextualised safety communication: Require operators to publish comparative safety data (AV vs. human driver benchmarks) not just absolute incident counts
  3. RAO disclosure regime: Passenger right to know whether their journey involved remote operator intervention; disclosure of RAO location and qualification standard
  4. Community engagement prior to deployment: Structured pre-deployment consultation — not notification — with affected communities; independent of operator-run engagement
  5. Incident reporting in plain language: Public incident reports in accessible, non-technical language within defined timeframes

Sources

Document Status: Compiled April 2026. Trust survey data current Q1 2026; California DMV data current November 2025.

Thread