■ SAFE RISK ■ Transportation & Logistics
No — but only because nobody would bother. This trade is being ended by regulation, cycle and auto rickshaws, and app-based transport, all of which arrived long before anything called AI.
“In Kolkata, this is tradition and livelihood. Automation isn't the issue. Everything else is.”
Our AI replacement risk score — how we score jobs
Hand-pulled rickshaw work in Kolkata is brutal and specific: barefoot or in thin sandals, hauling a two-wheeled wooden carriage through congested lanes, waiting at fixed stands, negotiating fares verbally, and thriving in monsoon flooding when the streets are impassable to every motorized vehicle — which is precisely the niche that keeps the trade alive. Most pullers rent their rickshaw daily from an owner, live in shared quarters, send money home to Bihar, and work until their body stops allowing it. Earnings are cash, unrecorded, and thin.
Automation is a non-factor here, and pretending otherwise misreads the economics. Nobody is building a robot to do a job whose entire cost advantage is a human working for very little in streets too narrow, flooded and chaotic for autonomous systems. What has displaced pullers is far more prosaic: cycle rickshaws and auto rickshaws, motorbike taxis, the metro's expansion, and ride-hailing apps that undercut the stand-based fare negotiation. Municipal policy has repeatedly attempted to ban or phase out the hand-pulled rickshaw on humanitarian grounds, and licences have not been newly issued for decades in practice.
What keeps the remaining fleet moving is a handful of stubborn advantages. Flooded streets, lanes no car fits down, short school runs where families trust a known puller, elderly regulars, and heritage tourism. There is also a social contract: the trade is a survival occupation for migrants with no capital and no other entry point into the city's economy, which makes an abrupt ban a livelihood catastrophe rather than a reform. So our low risk score is honest but almost cruel — this job is not threatened by AI because it is already being ended by something else. The relevant question is not automation but whether displaced pullers get anything resembling a transition.
Automatability: our editorial assessment of current and near-term AI capability
Decline is already advanced and continues regardless of AI. Licence attrition, motorized alternatives and app-based transport have been shrinking the trade for decades, and most projections have hand-pulled rickshaws surviving mainly as heritage and flood-season transport into the 2030s. The determining factors are municipal policy and the ageing of the current pullers, not technology.
No, and it never made economic sense to try. Autonomous vehicles need mapped, orderly roads; the rickshaw's advantage is exactly the opposite environment — flooded, narrow, unmapped lanes. The trade is being ended by cycle and auto rickshaws, the metro, ride-hailing apps and municipal policy, all of which long predate current AI.
Monsoon flooding, lanes too narrow for cars, short trusted trips for children and elderly regulars, and heritage tourism. There is also inertia: the pullers are largely migrant workers with no capital, and abolishing their livelihood without an alternative has repeatedly proven politically harder than announcing a ban.
Historically, not much support. Many shift to cycle rickshaws, e-rickshaws, construction labour or delivery work, all of which require either capital or a licence they may struggle to obtain. Transition schemes have been announced repeatedly and implemented patchily. This is a policy failure, not a technological one, and it will decide whether the ending is managed or brutal.
Generally yes — higher daily earnings, dramatically less physical damage, and a longer working life. The barrier is upfront cost and licensing, which is precisely what a puller lacks. Where loan or subsidy schemes have actually reached drivers, the switch has been fast, which tells you the constraint was never willingness.