AI Rickshaw: A Secure, Geofenced, Solar-Assisted Autonomous Electric Rickshaw for Bangladesh and the Displacement Risk to the Lowest-Income Transport Workers

Bangladesh moves on rickshaws: a quarter of Dhaka's trips are made by cycle or battery-assisted rickshaw, and one to two million battery-run three-wheelers in the capital each support a driver from one of Bangladesh's lowest-income informal transport workforces. This article presents AI Rickshaw, a reference design for a driverless, app-summoned, solar-assisted electric rickshaw (AIR-1) that keeps the traditional three-wheel form, and uses it as a Bangladesh-specific case study in embodied artificial intelligence. Four contributions are made. First, a reference architecture adapted to lane-less, low-speed, high-density traffic. Second, software-enforced traffic compliance: geofenced operation in which highways, restricted roads, wrong-way movement and out-of-zone trips cannot be planned or executed, so that regulation becomes executable rather than merely instructive. Third, an integrated cyber-physical security architecture for theft, tampering, unauthorised control and passenger emergencies, in which a hard-wired safety microcontroller is the sole validated actuator gateway, so that compromise of the back end does not directly confer actuator authority. Fourth, an argument, with scenario evidence, that embodied AI, the combination of perception, robotics, electric actuation and digital platforms, can threaten the lowest-income manual occupations and not only knowledge work: under the base adoption scenario, substantial displacement of Dhaka's drivers could occur by 2041, and a just-transition framework of a per-vehicle-day levy, garage franchising, microfinance co-ownership and evidence-gated deployment is proposed. Solar energy is treated as assisting, not primary, supply. No physical prototype has yet been built; every quantitative value carries its evidence type, and the next phase is a hardware prototype validated in a private geofenced environment.