The robot team at Google DeepMind has announced three new advancements aimed at enhancing the decision-making capabilities of robots in complex environments. One of these advancements is a data collection system equipped with a "Robot Constitution," designed to ensure that robots do not harm humans while performing tasks. This system utilizes visual language models and large language models to provide robots with safety prompts by understanding the environment and task requirements. In field tests, Google deployed 53 robots, employing measures such as joint force thresholds and physical cutoff switches to ensure safety. These advancements are instrumental in driving the development of robotics technology, making it more adaptable to a variety of tasks and environments.