3 min read

Google's 'Robot Constitution': Ensuring the Safety of New AI Robots

Google's DeepMind robotics team has unveiled three groundbreaking advancements that promise to radically alter the robotics landscape. This isn't merely a technological leap; it's a pivotal shift in how we incorporate robotics into our daily lives, blending technological prowess with ethical foresight.

Shaping the future of advanced robotics
AutoRT, SARA-RT, and RT-Trajectory build on our historic Robotics Transformers work to help robots make decisions faster, and better understand and navigate their environments.

Firstly, the introduction of a "Robot Constitution," drawing inspiration from Isaac Asimov's "Three Laws of Robotics," marks a significant ethical advancement. DeepMind's system, embedded into a Large Language Model (LLM), ensures that even mundane tasks, like an office robot fetching printer paper, are performed without compromising human safety. This isn't just a technological step; it's a moral one, addressing a long-standing concern in robotics: ensuring our life-easing machines don't inadvertently turn hazardous.

Secondly, Google's AutoRT system heralds a monumental advancement in robotic perception and interaction. Marrying a Visual Language Model (VLM) with an LLM, these robots can now adapt and make informed decisions in new environments. This isn't just an upgrade; it's a paradigm shift, redefining our understanding of robotic capabilities.

AutoRT follows these four steps for each task.

Thirdly, the implementation of safety measures, like automatic shutdowns and physical kill switches, during the extensive testing of 53 AutoRT robots over 77,000 trials, isn't just about refining technology. It's about building trust and reliability in systems that may soon become integral to our workplaces.

These advancements, including the SARA-RT neural network and RT-Trajectory technology, signify a new era in robotics: safer, more efficient, and more seamlessly integrated into our lives. While fully autonomous, drink-serving or pillow-fluffing robots remain a distant dream, the path carved by innovations like AutoRT is leading us toward a future where such capabilities are not just imagined but anticipated.


Today's announcement about Google's "Robot Constitution" provides a striking contrast to the diverse approaches to robotics in the United States and Japan.

Japan, a pioneer in the concept of embodied AI, has focused on creating robots for companionship and practical applications like caregiving, firefighting, and physical therapy. Meanwhile, the West has often viewed robots through a lens of skepticism, as job threats or dehumanizing forces.

Cultural perceptions play a significant role in this contrast. In Western pop culture, robots are often depicted as menacing, like the Terminator. In Japan, however, robots are seen more positively as contributors to society. This difference stems from Japan's post-WWII recovery, where technology and robotics were pivotal in national rebuilding, giving rise to friendly, human-like robots like Astro Boy.

Astro Boy, created by Osamu Tezuka in 1951, embodies this nurturing perception of machines in Japan. Unlike the Western story of Pinocchio, where a boy becomes a thing, Astro Boy is about a thing becoming a real boy. Professor Tenma, grieving over his son Tobio's loss, creates Astro Boy, who, with his powers, fights for good, symbolizing Japan's long-standing commitment to harmoniously integrating AI and robotics into everyday life.

Nevertheless, the strides made by Google's DeepMind are not just technical milestones. They represent a shift in our collective approach to robotics, blending technological innovation with ethical and cultural considerations. As we enter this new era, it's crucial to recognize that the future of robotics is as much about the impact these machines have on our world as it is about their capabilities.