​Boston Dynamics and Google DeepMind Teach Spot to Reason​

2026-04-14 · Source: IEEE Spectrum · Field: Technology & Digital — Robotics & Autonomous Systems, Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Intermediate, short

Summary

Boston Dynamics has integrated Google DeepMind's Gemini Robotics-ER 1.6, a high-level embodied reasoning model, into its quadruped robot, Spot. This enhancement aims to improve Spot's usability and intelligence for complex tasks, particularly in industrial inspection applications. The updated Spot can now autonomously identify dangerous debris or spills, read complex gauges, and utilize vision-language-action models for environmental understanding. While the AI model enhances reasoning, challenges remain in aligning robot understanding with human intuition, especially concerning object manipulation and safety. Gemini Robotics-ER 1.6 currently relies solely on visual data, with future plans to incorporate semantic safety models and leverage customer data for training, aiming for a reliability threshold above 80% to ensure commercial viability and operator trust.

Key takeaway

For Directors of AI/ML evaluating embodied AI solutions for industrial applications, Boston Dynamics' integration of Gemini Robotics-ER 1.6 into Spot signals a significant step towards commercially viable autonomous inspection. Your teams should consider piloting Spot for tasks requiring reliable environmental understanding and automated hazard detection, but be mindful of current limitations in physical interaction and data requirements for advanced manipulation. Focus on applications where the robot's >80% reliability threshold provides clear operational value.

Key insights

Integrating advanced AI reasoning models into robots enhances autonomy for complex real-world tasks.

Principles

• Robot understanding must align with human intuition.
• Commercial viability requires high reliability (>80%).
• Physical data is crucial for training robust robot models.

Method

Gemini Robotics-ER 1.6 uses vision-only data for high-level reasoning and includes success detection via multiple camera angles to confirm object grasps.

In practice

• Deploy Spot for autonomous industrial facility inspection.
• Utilize Spot for reading complex gauges and identifying spills.
• Participate in beta programs for new AI capabilities.

Topics

• Boston Dynamics Spot
• Google DeepMind Gemini Robotics-ER 1.6
• Embodied AI
• Robotic Inspection
• Robot Reasoning

Best for: AI Scientist, Research Scientist, AI Product Manager, Robotics Engineer, AI Engineer, Director of AI/ML

Related on AIssential

• Boston Dynamics’ robot dog now reads gauges and thermometers with Google's AI — Google DeepMind's Gemini Robotics-ER 1.6 model significantly enhances robot perception and safety for complex industrial inspections.
• Google Deepmind's Gemini Robotics models will power Boston Dynamics' Atlas for industrial tasks — Google Deepmind's Gemini Robotics models will power Boston Dynamics' Atlas robots for industrial automation.
• Google Partners With Agile Robots in Latest AI Robotics Push — Google DeepMind is integrating Gemini AI models into Agile Robots' industrial hardware to enhance real-world robotic capabilities.
• Google Deepmind's Gemini Robotics-ER 1.6 gives robots a sharper brain for planning and perception — Gemini Robotics-ER 1.6 enhances robot autonomy through improved perception, planning, and instrument reading capabilities.
• Gemini Robotics-ER 1.6: Powering real-world robotics tasks through enhanced embodied reasoning — Gemini Robotics-ER 1.6 enhances robot autonomy through advanced spatial reasoning, multi-view understanding, and instrument reading.
• The Internet can't stop watching Figure AI's humanoid robots handling packages — Humanoid robot demonstrations, even impressive ones, offer narrow views of real-world capabilities and often face scrutiny regarding true autonomy.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by IEEE Spectrum.