When robots do not follow the script. Incident with a robot in a restaurant in California, which reminds of the real risks of AI.

A recent incident at a Haidilao restaurant in California revealed the practical risks of AI systems in everyday environments. A humanoid robot, AgiBot X2, began to move erratically during a presentation, causing damage and requiring employee intervention to stop it. Although the company stated that the robot did not malfunction but was affected by the confined space, the incident raises questions about safety protocols and control systems in real-life situations. Additionally, the incident highlights the importance of rapid and effective human oversight systems, as well as the simple design of robots, which can reduce risks. As AI advances in operational environments, the focus must shift from capability to reliability, as the greatest risks are often practical and immediate. This situation is a wake-up call for companies adopting AI and robotics, emphasizing the need for testing in real conditions and safer design solutions.