A laboratory at Shanghai Jiao Tong University in China can design and run molecular biology experiments with almost no human intervention. Described on November 4, 2026 in Nature, the AutoDNA system uses a team of AI agents and robotic arms to carry out work from start to finish from a simple operator prompt.
How the autonomous lab works
Instead of adding AI to an existing system, AutoDNA was built from the start with AI at its centre. Robotic arms move samples between more than 20 types of instruments, and a single project can involve thousands of operations, including synthesizing and sequencing DNA.
According to Nature, after the initial prompt, human hands are needed only to refresh reagents and service the hardware. Haibo Chen, a computer scientist at Shanghai Jiao Tong University and a co-author of the study, explains that in many current systems experts still have to configure the instruments, while AI usually handles only one part of the process.
The tests: mutation, a message in DNA and storage
In the reported tests, the system detected a cancer-linked mutation in a DNA sample. In another experiment, the agents encoded the sentence “Intelligence is the ability to adapt to change” in DNA, using the four nucleotide bases A, C, G and T, and then read the content back. Coverage of the study reports that this DNA storage demonstration used 78 strands and took about 203 hours.
Milad Abolhasani, a chemical engineer at North Carolina State University, told Nature that the system stands out for its degree of autonomy and its ability to incorporate diverse equipment, adding: “The future of science is going to involve a lot of autonomous labs.”
What AutoDNA is not yet
The result does not mean labs can now run indefinitely with no human presence, nor that DNA is ready for large-scale archival storage. The mutation detection was a system-validation demonstration, not a new clinical test, and the recovered message was a controlled demonstration, not proof of industrial-scale storage. The advance lies elsewhere: showing that AI agents can orchestrate a complete molecular biology workflow across several instruments, rather than automating only one isolated step.
Image credit: Nature, showing DNA sequencing peaks displayed on a screen.
Sources: Nature — AI agents run this automated biology lab; Nature — An AI-native experimental laboratory for autonomous biomolecular operations.
Transparency: this article was produced with AI assistance, cross-checked against public posts on X and internet sources, and editorially reviewed before publication. Check the original sources listed below to confirm the details.
Muse
