DeepMind's New AI Enables Robots to Complete Tasks Via Web Search
Google DeepMind has announced that its enhanced AI models now empower robots to handle more intricate tasks and even utilize web resources for assistance. During a press briefing, Carolina Parada, Google DeepMind's head of robotics, explained that the company's latest AI models collaborate to enable robots to "plan multiple steps in advance" before executing actions in the physical environment.
The system is driven by the newly released Gemini Robotics 1.5 and the embodied reasoning model, Gemini Robotics-ER 1.5, which are upgraded versions of the AI models Google DeepMind unveiled in March. Robots are no longer limited to performing individual tasks like folding a sheet of paper or unzipping a bag. They can now undertake activities such as sorting laundry by dark and light colors, packing a suitcase according to London's current weather, and assisting with waste, compost, and recycling separation based on location-specific guidelines retrieved from web searches.
"Previous models excelled at carrying out individual instructions with high versatility," Parada noted. "With this update, we're shifting from single-step instructions to genuine comprehension and problem-solving for physical tasks."
To accomplish this, robots employ the upgraded Gemini Robotics-ER 1.5 model to interpret their surroundings and leverage digital tools like Google Search for additional information. Gemini Robotics-ER 1.5 then converts these findings into plain-language instructions for Gemini Robotics 1.5, allowing the robot to utilize the model's visual and linguistic capabilities to execute each step.

Google DeepMind's AI models collaborate to assist robots in planning and performing tasks. Image: GoogleRelated
- Google DeepMind's optimized AI model operates directly on robots
Furthermore, Google DeepMind revealed that Gemini Robotics 1.5 enables robots to "learn" from one another, regardless of their hardware configurations. The company found that tasks demonstrated on the ALOHA2 robot—a system with two mechanical arms—can be directly performed on the bi-arm Franka robot and Apptronik's humanoid robot, Apollo. "This provides two key advantages: first, we can control highly diverse robots, including humanoids, using a single model," stated Google DeepMind software engineer Kanishka Rao during the briefing. "Second, skills acquired by one robot can now be transferred to another."
As part of this rollout, Google DeepMind is making Gemini Robotics-ER 1.5 available to developers through the Gemini API in Google AI Studio, while Gemini Robotics 1.5 remains accessible only to selected partners.
Related article
Google Tests Remy AI Agent for Gemini as Focus Shifts to User Control
According to Business Insider, Google is testing Remy, a new AI personal agent for Gemini. This tool aims to execute tasks on behalf of users, streamlining both professional workflows and daily routines.Currently, Remy is undergoing testing in an int
Ollie bets privacy focus to win AI assistant race
To be genuinely helpful, an AI assistant must understand its user deeply. Ollie, a personal assistant designed for daily life, operates on the premise that this doesn’t require surrendering your data or compromising your privacy.While certain enterpr
How AI LIVE: London Will Explore AI & Industrial Automation
The summit will convene C-suite executives from around the globe to address pressing challenges in global industries, ranging from AI-driven disruption to economic volatility.AI LIVE: The London Summit will gather over 2,000 international leaders und
Related Special Topic Recommendations
Comments (1)
0/500
Google DeepMind has announced that its enhanced AI models now empower robots to handle more intricate tasks and even utilize web resources for assistance. During a press briefing, Carolina Parada, Google DeepMind's head of robotics, explained that the company's latest AI models collaborate to enable robots to "plan multiple steps in advance" before executing actions in the physical environment.
The system is driven by the newly released Gemini Robotics 1.5 and the embodied reasoning model, Gemini Robotics-ER 1.5, which are upgraded versions of the AI models Google DeepMind unveiled in March. Robots are no longer limited to performing individual tasks like folding a sheet of paper or unzipping a bag. They can now undertake activities such as sorting laundry by dark and light colors, packing a suitcase according to London's current weather, and assisting with waste, compost, and recycling separation based on location-specific guidelines retrieved from web searches.
"Previous models excelled at carrying out individual instructions with high versatility," Parada noted. "With this update, we're shifting from single-step instructions to genuine comprehension and problem-solving for physical tasks."
To accomplish this, robots employ the upgraded Gemini Robotics-ER 1.5 model to interpret their surroundings and leverage digital tools like Google Search for additional information. Gemini Robotics-ER 1.5 then converts these findings into plain-language instructions for Gemini Robotics 1.5, allowing the robot to utilize the model's visual and linguistic capabilities to execute each step.

Related
- Google DeepMind's optimized AI model operates directly on robots
Furthermore, Google DeepMind revealed that Gemini Robotics 1.5 enables robots to "learn" from one another, regardless of their hardware configurations. The company found that tasks demonstrated on the ALOHA2 robot—a system with two mechanical arms—can be directly performed on the bi-arm Franka robot and Apptronik's humanoid robot, Apollo. "This provides two key advantages: first, we can control highly diverse robots, including humanoids, using a single model," stated Google DeepMind software engineer Kanishka Rao during the briefing. "Second, skills acquired by one robot can now be transferred to another."
As part of this rollout, Google DeepMind is making Gemini Robotics-ER 1.5 available to developers through the Gemini API in Google AI Studio, while Gemini Robotics 1.5 remains accessible only to selected partners.
Ollie bets privacy focus to win AI assistant race
To be genuinely helpful, an AI assistant must understand its user deeply. Ollie, a personal assistant designed for daily life, operates on the premise that this doesn’t require surrendering your data or compromising your privacy.While certain enterpr
How AI LIVE: London Will Explore AI & Industrial Automation
The summit will convene C-suite executives from around the globe to address pressing challenges in global industries, ranging from AI-driven disruption to economic volatility.AI LIVE: The London Summit will gather over 2,000 international leaders und





Home






