Toward a 7×24 unmanned smart lab
Going forward, BioAgents will no longer improve through text-based training alone. Instead, it continuously accumulates real-world research tasks, automated operations, expert validation, failures, and wet-lab feedback through Physical AI’s experimental loop. This vast corpus of physical experiment data will enable BioAgent to develop integrated inference, execution, and validation capabilities that will ultimately power 7×24 unmanned intelligent laboratories.
Built on real lab experience
MGI’s AI exploration dates back to 2019. In 2025, a team led by MGI’s Chief AI Officer Dr. Yang Meng, in collaboration with Professor Nattiya Hirankan from Chulalongkorn University, published a paper. natural biomedical engineering “PrimeGen” is a dry/wet collaborative multi-agent system that integrates primer design, experimental validation, and automated workstation execution into a closed-loop workflow.
The new Physical AI initiative builds on MGI’s unique hardware-native advantages, with tight integration across MGI’s automation platforms and real-world deployment expertise from more than 3,800 users worldwide.
“This reflects a different path than pure computing competition. While major AI companies rely on scale computing to drive the capabilities of general-purpose models, we are taking a different approach. Through agent scaling and closed-loop data engineering, we organize real-world tasks, device constraints, expert feedback, and wet lab results into a training ground where AI continually evolves,” said Yan, currently CEO of Genoria AI. Dr. Meng said.
About Genoria AI
Genoria AI, a subsidiary of MGI Tech, is an AI4S company building self-evolving labs to scale agent discovery. For more information, please visit genoria.ai/en.
About Shanghai Artificial Intelligence Institute
The Shanghai AI Institute was officially unveiled at the World AI Conference (WAIC) in July 2020, and was positioned as a new national-level research institution. Our vision is to build a world-class AI laboratory, making pioneering contributions with unique theories and key technologies.
For more information, please visit shlab.org.cn.
