Anthropic's New Hardware Standard Lets AI Agents Control the Physical World
The framework lets trusted partners test AI control of lab and factory equipment before Anthropic opens it to wider use.
- On Thursday, Anthropic released a research preview of the Model Hardware Standard , a framework enabling AI agents to operate physical devices like microscopes and robotic arms in scientific research and manufacturing.
- Alek Kemeny, a quantum physicist who co-led the project, said "the impetus is wanting to accelerate science" by closing gaps between data analysis and experimental execution. The standard automates complex engineering tasks, eliminating the need for bespoke code across multiple robotic systems.
- The model-agnostic framework operates on any device with a programmable interface, allowing agents to execute autonomous workflows. Testing partners include Amazon Web Services, Danaher, Hugging Face, and Raspberry Pi, with a waitlist now open to others.
- Elizabeth Kelly, Anthropic's head of beneficial deployments, told CNBC the framework offers "huge benefits here for enterprise and for industry," while including safety guardrails like limiting robot speed and angles to prevent mishaps.
- Europe's Machinery Regulation 2023/1230, taking effect January 20, 2027, will cover AI-based safety functions, potentially classifying MHS files for high-risk categories as regulated safety components requiring formal compliance.
44 Articles
44 Articles
A startup is creating an open standard for neural networks. Anthropic has unveiled an experimental standard, MHS, that allows AI agents to interact with physical equipment. This could enable autonomous models to independently conduct experiments, control devices, and perform tasks without complex software, according to RBC-Ukraine, citing Ars Technica. Anthropic explains that MHS creates a common interface and data exchange format, allowing disp…
System allows to control scientific and engineering equipment for experiments and brand launch of the company's first tool to operate in the physical world
Anthropic prepares a common way to connect AI with robots, microscopes and other machines without programming each device from scratch
Coverage Details
Bias Distribution
- 40% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
