Google’s Gemini Empowers Robots

⊲ WIRED ⊳ Google DeepMind has introduced a new AI model called Gemini Robotics, which integrates language, vision, and physical actions to enhance robotic capabilities. This model enables robots to perform tasks such as folding paper and handing over objects based on verbal instructions. Simulation supports the training of the robots, enabling them to refine their skills before real-world deployment, accelerating learning and reduces risks. With generalised behaviour across varied hardware, the model allows researchers to train their own robots. However, challenges remain regarding safety, prompting Google to establish an ASIMOV benchmark to assess potential risks in robot behaviour. The technology is still in early stages and may take years to develop further.
⊲ Image – Google ⊳