When Amazon launched its cashier-less stores, Just Walk Out, the innovation was hailed as the future of retail. Customers enter stores using palm recognition connected to a credit card. Cameras then capture images of what customers take. Then, artificial intelligence is responsible for analyzing them and billing them automatically. At the beginning of April, however, Amazon announced that it was removing the technology from its American Amazon Fresh grocery stores – even if it is maintaining it in its Amazon Go stores. The innovation was then criticized. Critics suggested that Amazon was using a thousand workers in India to carry out the task supposedly carried out by artificial intelligence. In reality, these workers are responsible for comparing actual purchases with what was analyzed by the algorithm in order to retrain it. A very classic process in machine learning.
The use of “mechanical Turks” has often been denounced. The expression refers to this famous hoax constructed at the end of the 18th century where an alleged chess-playing automaton was in fact animated by a human. Tiny Mile, which runs a home food delivery robot service, relies on pilots in the Philippines, while Kiwi’s robots, used on some U.S. college campuses, use workers in Colombia earning less than 2 dollars an hour to help them make deliveries. In the exceptional register, we sometimes accept that a surgeon takes remote control of an automated device to carry out an intervention.
But what is new today is the assumed provision of a daily contact service by a remote human. In New York, in the fast food chain of Japanese origin Sansan Chicken, you will be surprised to see that the employee who takes your order is located in his home in the Philippines and interacts with customers via Zoom. He is employed by the company Happy Cashier, which allows restaurants to outsource their workforce overseas. The phenomenon is not as new as it seems. Canadian fast food chain Freshii was at the heart of a controversy in 2022 when it was revealed that its remote cashiers based in Nicaragua earned $3.75 per hour compared to $16 for a physical cashier.
Control hazards
Companies tend to prefer this experience to a pure automaton because it improves customer satisfaction and allows for control over hazards. Having an exchange with a human, even remotely, arouses more empathy and therefore limits the risk of annoying the customer. Retailers from Dollar General to Walmart and Costco are also rethinking the use of self-checkouts as they find they lead to higher merchandise losses due to customer errors and shoplifting. .
Reception agents at the entrance to buildings are another use case. This can be a person who validates the opening of the door remotely or who welcomes visitors via a large screen. The price of this service is often a third of that of a full-time person. The Virtual Doorman company equips 400 residential and tertiary buildings in New York alone. But this service is also distributed in emerging countries such as Argentina or Uruguay, as the entrepreneur Pieter Levels (Nomad List) was surprised recently.
This remote work is made possible by the ever-increasing fluidity of telecommunications and the increased quality of cameras. This is what allows us to imagine the remote realization of other local service professions. This is the bet of the German start-up Vay which has just launched its valet parking service in Las Vegas. A simple click and an electric car comes to you but no one is sitting in the driver’s seat. Because the driver is remotely in a simulator with a steering wheel, pedals, brakes and three screens. You drive the car while you get to your destination and just have to get out to let the valet take control while you park the vehicle. The service costs 30 cents per minute to drive the car compared to $25 for an on-site valet. A whole new kind of teleworking.
Robin Rivaton is Managing Director of Stonal and member of the Scientific Council of the Foundation for Political Innovation (Fondapol)
.