Anthropic Unveils Standard to Connect AI With Robots and Industrial Hardware -

Anthropic PBC has introduced a new software standard designed to help artificial intelligence (AI) assistants interact more effectively and safely with robots, scientific equipment and manufacturing hardware.

The AI company has released its Model Hardware Standard (MHS) as a research preview, allowing developers and industry stakeholders to test how its Claude AI assistant can work with physical equipment before the framework is made more widely available.

MHS is designed to provide AI systems with structured information about how hardware operates, reducing reliance on paper manuals or specialised knowledge held by a limited number of experts. For example, a manufacturer could use the standard to define how an AI system can safely control a factory robot arm by setting limits on its speed, movement and operating angles.

The launch comes as interest grows in applying AI to robotics, manufacturing and scientific research. Anthropic said stakeholders in these sectors can join a waitlist to test the framework, which it eventually plans to open-source, allowing hardware manufacturers to create their own MHS specifications. The company did not provide a timeline for a broader release.

Jonah Cool, a partnerships and deployment executive at Anthropic’s life sciences division, said the framework could help scientists operate complex equipment that might otherwise require specialised technical expertise.

Anthropic compared MHS to its Model Context Protocol (MCP), a standard that helps AI assistants connect with third-party software and applications. According to the company, MHS aims to bring similar interoperability to the physical world, particularly in laboratories and facilities that rely on large numbers of devices.

The company has already tested the framework with industry stakeholders including Amazon Web Services, Danaher, Hugging Face and Raspberry Pi. In one demonstration, Anthropic said a scientist was able to provide Claude with an experimental design, after which the AI assistant carried out the experiment using hardware configured with the MHS specification.

The initiative reflects the growing push by AI companies to move generative AI beyond software applications and into real-world environments, including robotics, manufacturing and scientific research. – TS/ERMD

READ MORE