• by Laurie Sullivan , Staff Writer @lauriesullivan, Yesterday

Google DeepMind has released a Gemini Robotics artificial intelligence (AI) model that can run on a local device. It is designed for controlling robots, but its capabilities have implications for advertising, particularly performance advertising.

Gemini Robotics On-Device, as creators call it, runs with a voice language action (VLA) model that can make robots perform a range of tasks in physical environments. The AI model functions are independent from a data network, making it more useful for applications that are latency sensitive. 

The model comes from a family of AI technology designed to bring advanced capabilities to robots. It adheres to natural-language instructions and performs complex tasks such as unzipping bags or folding clothes.

Gemini Robotics On-Device drives robotic actions in the physical world that open possibilities for AI-powered physical advertising such as digital displays that can react to people, showing personalized ads based on their demographics or past interactions.

One example would be a robot in an auto parts store that has a screen. The consumer could ask for a specific auto part, and the robot could show the person on the screen how to fix it depending on the make and model of their car. In my opinion, the examples are unlimited.

The technology also would have the ability to offer personalized product recommendations or provide guidance for customers. Imagine the robot as a clothing sales associate or concierge in a retail store answering consumer questions and completing purchases. Those who have spent time interacting with the robots at the Sphere in Las Vegas can clearly see how companies can use robots to improve performance advertising.

Carolina Parada, senior director and head of robotics at Google DeepMind, announced the release of Gemini Robotics On-Device in a blog post on Tuesday.

The company also made the model accessible through a Gemini Robotics software development kit (SDK) for those participating in a beta program. 

"Our model quickly adapts to new tasks, with as few as 50 to 100 demonstrations — indicating how well this on-device model can generalize its foundational knowledge to new tasks," Parada wrote in a post. 

The robots will become key in improving customer experience, and understanding behavior will be necessary.

Robots equipped with sensors and AI gather data on consumer behavior and preferences. They have cameras that record consumer interactions with products and displays, and could keep an eye out for theft.

Retailers can create personalized recommendations and tailor marketing messages that speak with individuals.