The robotaxi has crossed the line from demonstration to utility. In a growing list of cities, residents summon driverless cars the way they summon any ride, and the vehicles — patient, literal-minded, festooned with lidar — complete millions of paid trips. Expansion announcements now arrive quarterly: new cities, airport pickups, freeway routes, late-night hours. The debate has matured accordingly. The question is no longer whether a car can drive itself across town, but how society should evaluate machines that do so imperfectly in public.
The operators’ strongest card is data. Insurer-analysed comparisons of robotaxi fleets against human-driver benchmarks report substantially lower rates of injury-causing crashes per mile — figures in the range of 70 to 85 per cent fewer for the largest operator’s published analyses, drawn from tens of millions of rider-only miles. Regulators such as the US National Highway Traffic Safety Administration require standing crash reporting for automated systems, creating a public record no human fleet has ever had to match. Sceptics correctly note the comparisons are imperfect: robotaxis drive chosen routes in mapped cities, in decent weather, at modest speeds — a curated slice of driving, not its hardest chapters.
The Edge Cases That Remain
The failures, when they come, are alien rather than human: vehicles freezing before emergency scenes they cannot parse, creeping through flooded streets, misreading a pedestrian carrying something large. Cities have fought for — and partly won — better data access and incident protocols after early clashes over blocked lanes and confused first responders. Labour politics run alongside the safety file: driver organisations argue the industry is automating livelihoods without a transition plan, while operators counter with accessibility stories — blind riders, elderly passengers — that are genuinely hard to dismiss. Our transport coverage continues in our technology section.
Gradual, Then Sudden
The Economics Under the Safety Story
Safety may win the argument, but economics decides the map. A robotaxi’s cost per mile — vehicle, sensors, remote assistance, cleaning, insurance, depreciation — must fall below the human-driven alternative’s in each city it enters, and the industry’s expansion order follows that maths: dense, affluent, well-mapped, regulation-friendly. The vehicles are also quietly reshaping their host systems: taxi and ride-hail driver earnings in served zones, kerb space politics, and public transport ridership on the corridors they parallel all appear in city data now. The most credible forecasters expect a long hybrid era — humans and robots sharing the road and the market, sorted by trip type and neighbourhood — rather than a takeover. History agrees: technologies of transport replace each other the way the car replaced the horse, which took decades, happened unevenly, and was obvious only in retrospect.
Regulators are developing a new literacy in response: reading fleets. The permit regimes maturing in the leading cities no longer ask merely whether a service may operate, but require continuous reporting — disengagements, near-misses, blocked lanes, kerb behaviour — audited against the companies’ public claims. That data relationship is the real institutional innovation of the robotaxi era, and it cuts both ways: cities gain an evidence base for transport policy they never had from human drivers, while companies gain a regulatory moat, since only well-capitalised fleets can afford the reporting apparatus. The kerb, the crossing and the incident report are where this technology will actually be governed. The steering wheel was never the point.
The pattern of adoption looks like other transport technologies: expensive, geographically fenced, then suddenly ordinary. Costs per vehicle keep falling as sensor prices drop; the map keeps widening city by city. The safety argument will likely be settled not by any single study but by accumulation — mile upon audited mile, compared honestly against the human baseline of roughly one death per hundred million miles that societies have long treated as normal. The robotaxi’s deepest challenge to its critics is that the status quo it must beat is not perfection. It is us.
