Google’s Gemini Robotics AI Model Reaches Into the Physical World
1 min read
Summary
Search engine giant Google’s artificial intelligence (AI) unit DeepMind has created a new model called Gemini Robotics, which it says can control a range of robots to perform different tasks.
DeepMind demonstrated over a series of video calls how the robots could fold paper, pick up vegetables and gently put glasses away.
The robots are also able to learn and complete tasks in hundreds of scenarios not included in their original training.
However, experts have raised concerns that putting AI models in charge of physical robots could potentially be dangerous.
To mitigate this risk, the company has also announced a new benchmark for gauging the risks of AI-powered robots, called ASIMOV.
