Anthropic Tests a New Standard for Claude: Collaborating with Factory and Lab Hardware
Introduction
Anthropic has released the Model Hardware Standard (MHS) in research preview, marking a significant step forward in ensuring AI-human collaboration with physical machinery. This standard outlines how AI should interact with hardware, including safety functions, especially as regulations around AI-based safety are set to change dramatically in January 2027.
The Model Hardware Standard
The MHS is a specification that documents the functionality and operating limits of a piece of hardware. For instance, it can specify how an AI should control a robot arm, setting speed and angle restrictions. This way, AI models don’t need to rely on paper manuals or experts for instructions.
“In many cases, the science doesn’t happen because you can’t use the equipment,” noted Jonah Cool of Anthropic’s life sciences arm.
The MHS follows in the footsteps of the Model Context Protocol (MCP), which connects AI assistants to various services like Gmail, Calendar, and Slack.
A Test Group of Note
Amazon Web Services, Danaher, Hugging Face, and Raspberry Pi are among those testing the MHS. Interestingly, two of these companies have European origins: Raspberry Pi is based in the UK, while Hugging Face was founded by a French team.
Europe’s regulatory landscape is also evolving to accommodate AI-based safety functions, with new regulations coming into effect on January 20, 2027. This change will impact high-risk machinery and AI systems with self-evolving behavior.
Open-Sourcing for the Future
Once the research preview period ends, Anthropic plans to open-source the MHS framework. This development could mean that anyone creating hardware specifications for AI safety will need to adhere to a regulated standard, as defined by the new European regulations.
By Darius Popa
Technology enthusiast and intern at The Next Web, contributing to research-backed content and investigating new technologies and global events.
Interested in business and how the narrative and perception of technology is shaped.