Beijing has recently become the site of an experiment merging technology, public engagement, and live AI testing. A humanoid robot is now assisting customers at an automated kiosk, eliciting a range of reactions. While its service remains slow and limited, the engineers are less focused on the customer experience than on the hidden capabilities being explored beyond the lab.
The G1, created by Galbot, works at a public beverage kiosk. Unlike conventional machines, it doesn’t rely on a preset sequence of motions. Its system integrates real-time computer vision with a language model, enabling it to understand its surroundings, make decisions, and act autonomously.
A Mobile Base Designed for Adaptability
Although the robot has no legs, it moves on a mobile base built for stability and precise control. Equipped with cameras across its body and multiple sensors, it can recognize objects, assess their positions, and adapt its movements to subtle changes in its environment.
Such independence is a notable improvement over robots that require tracks, floor markings, or fully managed environments.
From an outsider’s perspective, the G1’s slow pace in grabbing a bottle or serving a drink might appear as a weakness. Yet for Galbot’s engineers, this deliberate slowness is crucial for training. Every movement produces important data that trains the system’s learning algorithms, helping it improve its adaptability.
Unlike machines built just for immediate efficiency, the G1 is designed to learn from its surroundings. The kiosk acts as a dynamic laboratory, where people move around, objects shift, lighting changes, and unexpected noises occur—conditions that are impossible to replicate entirely in controlled environments.
Although the G1 is currently handling a simple task, Galbot’s ultimate goal is far more ambitious. The robot was engineered for logistics, distribution centers, and dynamic object-handling environments.
Demonstrated Skills and Logistics Potential
In internal trials, it has already shown it can reorganize boxes, adjust movements when objects are unexpectedly displaced, and modify its approach as new obstacles appear. This demonstrates potential applications in warehouses, goods transportation, and last-mile delivery operations.
The Beijing kiosk isn’t meant to replace service workers. It serves primarily as a technological showcase and as a real-world endurance test for a versatile robot.
For now, the G1 may seem more of a novelty than a practical tool. Yet beneath its slow movements, a new generation of robots capable of perceiving, deciding, and acting autonomously is emerging—a subtle but crucial step toward a future where machines don’t just follow instructions, but learn from the world around them.
Read the original article on: Gizmodo
Read more:Can Digital Cadavers Replace Real Bodies in Medical Training