Anthropic has announced the Model Hardware Standard (MHS), a shared specification for AI agents to safely operate physical devices, as part of a research preview. The initiative aims to simplify the integration of hardware in labs and manufacturing settings by allowing AI agents to manage multiple instruments, such as microscopes, liquid handlers, and robotic arms, simultaneously. MHS was developed through a collaboration between Anthropic and HHMI Janelia Research Campus. The standard is designed to reduce the time required for integration from weeks or months to hours or minutes, making it easier for researchers and engineers to orchestrate autonomous experiments and workflows.
MHS introduces a standardized driver that translates between a computer’s operating system and a hardware device, using simple commands like 'read' or 'write' that any device can understand. This driver also helps AI agents learn about new devices by providing information about their characteristics, such as the weight of a robot arm, which is crucial for safe operation. The driver allows users to input this information in natural language, either manually or through an AI agent that interviews them about their hardware setup. This information is then used to generate a reference file that gives the agent all the details needed to operate the device.
The development of MHS began as a collaboration between Anthropic and HHMI Janelia Research Campus. The initiative was shared with a group of scientific research labs and advanced manufacturers to build safety evaluations and develop best practices for AI systems operating physical equipment, ahead of making the standard open source. MHS works with any device that has a programmable interface and is model-agnostic, allowing any agent harness to access it using standard protocols, such as the Model Context Protocol.
Source: anthropic
