Google DeepMind faces union talks while launching Gemini Robotics-ER 1.6 AI model

Google DeepMind is dealing with two major developments in 2026. Engineers at the AI research unit have begun unionization talks, seeking formal representation to influence the ethical deployment of the advanced systems they build. Management’s response has been described as minimal and obstructive, highlighting a dispute over internal governance rather than wages.

At the same time, DeepMind released Gemini Robotics-ER 1.6, the latest version of its embodied‑reasoning AI for physical systems. The model achieves up to 93% accuracy on industrial instrument‑reading tasks and improves spatial reasoning, multi‑view understanding, and success‑detection capabilities. Boston Dynamics has already integrated the model into its Spot robot, and developers can access it via the Gemini API and Google AI Studio.