- Anthropic introduced the Model Hardware Standard (MHS) for lab AI communication and coordination
- MHS links lab instruments and computers via a unified interface for diverse equipment collaboration
- Researchers can direct AI to control multiple instruments in experiments using plain language input
Anthropic has unveiled a new software framework that could give AI agents a direct line to the laboratory floor, allowing machines from different manufacturers to communicate, coordinate tasks and work together on experiments through a single interface, according to a Nature report.
Anthropic's Model Hardware Standard (MHS) creates a unified digital bridge linking lab computers and scientific instruments with AI systems, allowing them to interact and coordinate through a shared framework.
ALSO READ : Google Tests 'Call For Me', Letting Gemini AI Make Phone Calls On Users' Behalf
Also Read | Layoffs Are Rising In Tech, But These Skills Are Helping People Unlock Better Pay
According to the report, Anthropic built the framework alongside researchers at the Janelia Research Campus, part of the Howard Hughes Medical Institute. The initiative is aimed at solving a persistent challenge in laboratory automation: getting equipment from different manufacturers to work together seamlessly.
Scientific instruments are often built by different manufacturers and operate using their own software, programming languages and communication protocols. Connecting them can therefore require separate integrations for individual machines.
MHS aims to simplify that process by giving laboratory equipment a common interface through which it can communicate with other devices and AI systems.
How Anthropic's AI Lab Framework Works
Under the system, researchers can connect programmable laboratory equipment and describe its functions to an AI agent in plain language. Technical documentation, including equipment manuals, can also be supplied to help the AI understand how a particular instrument works.
Once the equipment is connected, researchers can describe the experiment they want to conduct. The AI agent can then coordinate the instruments required to carry out the workflow.
In one demonstration cited by Nature, a robotic arm moved a multi-well plate between different pieces of laboratory equipment. The plate was first transferred to a liquid-handling machine and then moved to another instrument for analysis.
ALSO READ : Zuckerberg Says Meta's New Muse Wearables Bring 'Personal Superintelligence' Within Reach
In a company statement issued on August 27, Anthropic technical staff member Alek Kemeny said the MHS framework serves as a digital link between laboratory computers and scientific instruments.
The framework is also being tested by researchers at Carnegie Mellon University in Pittsburgh.
Computational biologist Jose Lugo-Martinez and his PhD student Sina Barazandeh said an experiment that would traditionally take months to set up was completed in a matter of hours using the framework, according to Nature.
AI Could Control Multiple Lab Instruments
The technology is also being tested at Janelia, where neuroscientist Virginie Ruetten is using MHS in experiments involving zebrafish. Ruetten works with more than 20 instruments and seven software programmes.
The framework allows her to initiate experiments through a single interface rather than having to operate each system separately. Anthropic plans to make MHS freely available, with early access already provided to selected research groups.
The company's approach is not the first attempt to standardise communication between laboratory instruments.
The Standardisation in Lab Automation (SiLA) initiative has already developed common standards for laboratory equipment. However, MHS is aimed specifically at enabling AI agents to interact with and coordinate laboratory instruments, potentially adding an automation layer on top of existing laboratory systems.
Challenges Remain For Fully Automated Research
Despite the potential benefits, experts caution that connecting machines is only one part of automating scientific research.
Laboratory equipment often requires highly precise instructions. Cristian Ponce, CEO of lab automation company Tetsuwan Scientific, told Nature that robots need specific programming for tasks such as pipetting, including the viscosity of a liquid, the speed at which it is drawn and the precise position of the pipette.
That means AI systems will require detailed technical information about individual instruments before they can reliably operate them.
The development of MHS nevertheless points towards a broader shift in laboratory research, where AI agents could move beyond analysing scientific data and begin coordinating physical experiments.
If such systems can reliably communicate with and control diverse laboratory equipment, researchers could automate more complex workflows while reducing the time and effort required to set up experiments.
Essential Business Intelligence, Sharp Market Insights, Practical Personal Finance Advice, Daily Fuel, Gold and Silver Prices and Latest Stories — On NDTV Profit.
