Anthropic Unveils MHS Standard for AI-Controlled Machines

Anthropic has previewed the Model Hardware Standard (MHS), developed with medical research institute HHMI, to help AI agents control scientific instruments and industrial machines. The standard replaces device-specific APIs with a single unified interface, enabling AI agents like Claude to automate lab workflows, fix experiment errors, and generate instrument control scripts. Early adopters include HHMI, which built an AI agent to manage microscopes, and quantum computing startup QuEra. AWS, Hugging Face, and several robotics firms are integrating MHS into their products. Anthropic plans to release the standard under an open-source license after refining safety features with initial partners.