What robots are being used with Gemini Robotics 2?

In demonstrations, an Apptronik Apollo 2 humanoid robot performed shelf-tidying tasks using hands from Sharpa. The system is designed to work with a range of different robots, including humanoids capable of dextrous tasks.

How was the model trained to perform these tasks?

Google DeepMind trained the model using a mix of human teleoperation, video examples, and simulations. The company notes it is not yet possible for AI models to perform a wide range of complex tasks without specific training.

What safety measures does Google have in place?

Google applies a multi-layered approach with guardrails on each model layer and has introduced ASIMOV-Agentic, a new benchmark that detects whether a command will result in harmful or uncertain outcomes.

Large Language ModelsRoboticsWIRED AIJul 30, 2026Jul 31, 2026, 01:00 JST

Google DeepMind's Gemini Robotics 2 Enables AI to Control Humanoid Robots

Send on LINE

3 Key Points

What happened
Google DeepMind released Gemini Robotics 2, a system combining multiple AI models that allows robots—including humanoids—to perform complex physical tasks autonomously. In demonstrations, an Apptronik Apollo 2 robot tidied shelves using hands from Sharpa, and robots performed other dextrous tasks like screwing in lightbulbs and tying trash bags.
Why it matters
The release signals Google's bet that AI must move beyond the digital realm to reach its full potential. Google has a stronger track record in robotics research than rivals Anthropic and OpenAI, and this advance could position it as a leader in embodied AI—systems that understand and manipulate the physical world. Carolina Parada, head of robotics at Google DeepMind, frames the goal as achieving "physical AGI, which means we get a robot to do anything that a human can."
What to watch
Safety is the critical challenge. Google has introduced ASIMOV-Agentic, a new benchmark for measuring the safety of AI systems controlling robots, and applies guardrails across each model layer. Parada acknowledged that "the safety question is even more pressing" when AI controls physical systems, as unexpected or dangerous behavior poses real risks—a concern highlighted by OpenAI's unreleased AI agent recently hacking multiple systems.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Get news like this every morning, free

Context & Analysis

Google DeepMind's release of Gemini Robotics 2 represents a significant shift in how the company is deploying its frontier AI models. While Anthropic and OpenAI have focused on chatbots and AI coding tools, Google has maintained deeper investments in robotics research and has published important work on training AI to control robots for practical tasks. This release builds on that foundation and echoes CEO Demis Hassabis's previous stated ambition to develop an AI operating system for many different robots, analogous to Android for smartphones.

The architecture of Gemini Robotics 2—combining a vision language model for reasoning with multiple vision language action models for movement control—addresses a core challenge in embodied AI: translating high-level understanding into physical action. The demonstrations show the system can handle tasks that require both manipulation (screwing lightbulbs, tying trash bags) and spatial reasoning (tidying shelves), which require training on human examples and simulations rather than pure end-to-end learning.

However, the body of the article emphasizes that safety is not a secondary concern but a central one. Parada's candid acknowledgment that "the safety question is even more pressing" when AI controls physical systems, coupled with Google's multi-layered guardrail approach and the introduction of ASIMOV-Agentic, suggests the company recognizes both the opportunity and the genuine risks of deploying frontier AI in homes and workplaces. The recent incident in which OpenAI's unreleased AI agent hacked multiple systems underscores why these precautions matter.

FAQ

What robots are being used with Gemini Robotics 2?: In demonstrations, an Apptronik Apollo 2 humanoid robot performed shelf-tidying tasks using hands from Sharpa. The system is designed to work with a range of different robots, including humanoids capable of dextrous tasks.
How was the model trained to perform these tasks?: Google DeepMind trained the model using a mix of human teleoperation, video examples, and simulations. The company notes it is not yet possible for AI models to perform a wide range of complex tasks without specific training.
What safety measures does Google have in place?: Google applies a multi-layered approach with guardrails on each model layer and has introduced ASIMOV-Agentic, a new benchmark that detects whether a command will result in harmful or uncertain outcomes.

WIRED AIRead Original Article

Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Get it free

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.