Anthropic has introduced the Model Hardware Standard (MHS), a new specification designed to let AI agents safely operate equipment in the physical world. The framework allows agents to control multiple programmable devices, including microscopes, liquid handlers, robotic arms, lasers, and manufacturing equipment, rather than limiting their actions to software environments.
MHS creates a standardized interface between AI agents and hardware, allowing models to monitor equipment, sequence operations, adjust parameters in real time, and in some cases recover from hardware errors. Anthropic has demonstrated Claude using the system to adjust and calibrate a laser after observing the results through a camera.
The technology is currently available as a limited research preview to selected scientific laboratories and advanced manufacturers. Anthropic plans to use the testing period to develop safety evaluations and best practices before eventually making MHS open source.
