Anthropic opens AI tool that can independently conduct scientific experiments

Anthropic, the AI safety company, has introduced a Model Hardware Standard that allows artificial intelligence agents to interface directly with scientific equipment and conduct experiments with minimal human intervention. The research preview enables systems such as Claude to control laboratory instruments, automate experimental procedures, and coordinate complex workflows across multiple lab environments.

The technology represents a notable advancement in expanding AI capabilities beyond text-based interactions. Rather than simply providing recommendations or analyses, these AI agents can now execute physical tasks in laboratory settings, potentially accelerating scientific research and reducing the time required for routine experimental work.

Through the hardware standard framework, AI systems can manage data collection, equipment calibration, and workflow coordination—functions previously requiring direct human oversight. The development opens possibilities for more efficient laboratory operations, particularly in fields requiring repetitive experimental protocols or continuous monitoring of conditions.

The research preview indicates Anthropic's strategic push toward practical AI applications in scientific domains. As AI systems become capable of independent laboratory work, questions around oversight, safety protocols, and verification procedures remain central considerations for widespread adoption in research institutions across Asia and globally.