Why deploying autonomous vehicles and AI cameras without solving discovery liability, union transition pathways, and the kerb-to-front-door accessibility gap will derail public trust.

By Michael Castleman · michaelcastleman.me · November 2026 · 7 min read

Executive Summary

Technology is the easy half of autonomous transport. The hard half is governing the six systemic blind spots that emerge when every vehicle records 360-degree LiDAR/video and human drivers are no longer in the cab to assist vulnerable passengers.

Key Verified Metrics

  • Risk #1: Data Liability: 'If we hold the 360° video/LiDAR of a hazard and didn't look at it before a crash, courts treat it as negligence'
  • Risk #2: Union Co-Design: Bringing RTBU, TWU, ASU & maritime unions onto the transition journey before procurement, not after
  • Risk #4: The Last 15 Metres: A driverless pod can drop a blind passenger at the kerb — but who guides them across the footpath to their front door?
  • +20% to +40% VKT Risk: Potential jump in 'zombie miles' from empty robotaxis and single-parcel pods without dynamic kerb pricing

1. Holding Data Risk ('If We Have It, We Must Have Looked at It') & AI Traceability

Autonomous taxis,council garbage trucks, and iPAVE scanners record continuous 360-degree video and LiDAR of every road defect, tree overhang, and near-miss. If an agency stores petabytes of raw footage showing a failing guardrail or pothole and fails to act before a fatal collision, plaintiff lawyers can subpoena that unreviewed data as proof of negligence. Agencies must enforce automated edge-triage, immediate work-order slotting, and strict edge-redaction/deletion windows — alongside ISO/PAS 8800 immutable flight-recorder logs for end-to-end neural AI decisions.

2. How Does the Blind or Disabled Passenger Get from the Kerb to the Front Door?

A step-free autonomous pod can lower its suspension at the kerb, open its sliding door, and announce the stop. But for a blind passenger, a frail older citizen, or a traveler with cognitive disability, a human taxi driver currently provides crucial 'last 15 metres' physical and verbal assistance across busy footpaths, around café chairs, and up to the front door. Removing the human driver requires mandating 'Human-Assisted Mobility Tiers', station/precinct concierges, and NaviLens/UWB 3D spatial audio doorway beacons so no citizen is left stranded at the kerb.

3. Union Co-Design, Induced 'Zombie Mile' Congestion & Grid Megawatt Spikes

Successful automation requires transparent retraining agreements with transport unions mapping every displaced driving role into remote tele-ops, incident response, or depot robotics — paired with dynamic zero-occupant road pricing to prevent empty repositioning pods from choking CBD streets, and V2G depot microgrids to buffer 6pm summer megawatt charging peaks.