Boston Dynamics’ robot dog now reads gauges and thermometers with Google's AI

2026-04-15 · Source: AI - Ars Technica · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Emerging Technologies & Innovation · Depth: Intermediate, quick

Summary

Boston Dynamics' Spot robot can now accurately read analog instruments like gauges and thermometers in industrial facilities, thanks to Google DeepMind's new Gemini Robotics-ER 1.6 AI model, announced April 14. This model enhances "embodied reasoning" for robots interacting with physical environments, acting as a high-level reasoning model for task planning and execution. It significantly boosts instrument reading accuracy from 23% (Gemini Robotics-ER 1.5) to 98% by incorporating "agentic vision," a feature from Google's Gemini 3.0 Flash model. The Gemini Robotics-ER 1.6 model also improves multi-view reasoning, reduces "hallucination" problems in object identification, and is described as Google's "safest robotics model yet," with enhanced adherence to physical safety constraints and better perception of human injury risk.

Key takeaway

For Computer Vision Engineers developing autonomous inspection systems, the Gemini Robotics-ER 1.6 model's 98% accuracy in reading analog instruments and improved safety features suggest a significant leap in deployable robotic capabilities. You should evaluate integrating agentic vision models to enhance visual reasoning and reduce errors in complex industrial environments, particularly where human safety is a critical concern.

Key insights

Google DeepMind's Gemini Robotics-ER 1.6 model significantly enhances robot perception and safety for complex industrial inspections.

Principles

• Agentic vision improves visual reasoning.
• Multi-view reasoning enhances environmental understanding.
• Safety constraints are critical for robot deployment.

Method

The Gemini Robotics-ER 1.6 model uses "agentic vision" to combine visual reasoning with code execution, creating a "visual scratchpad" for inspecting and manipulating images, and employs a pointing process for complex visual tasks.

In practice

• Deploy robots for analog instrument reading.
• Utilize multi-camera systems for environmental context.
• Prioritize AI models with advanced safety features.

Topics

• Gemini Robotics-ER 1.6
• Boston Dynamics Spot
• Industrial Inspection
• Agentic Vision
• Embodied Reasoning

Best for: Computer Vision Engineer, Research Scientist, Robotics Engineer, AI Engineer, AI Scientist

Related on AIssential

• 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.
• ​Boston Dynamics and Google DeepMind Teach Spot to Reason​ — Integrating advanced AI reasoning models into robots enhances autonomy for complex real-world tasks.
• 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 enhances reasoning to help robots navigate real-world tasks. — Gemini Robotics-ER 1.6 enhances robot autonomy through improved spatial reasoning and new instrument reading capabilities.
• Google Gemini Is Taking Control of Humanoid Robots on Auto Factory Floors — Multimodal AI like Gemini can power general-purpose robots, enhancing physical intelligence and expanding applications.
• 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.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI - Ars Technica.