NEW DELHI: Walking into a laboratory where experiments are conducted without human hands may sound like science fiction, but researchers at IIT Delhi have made it a reality.

In a study published in Nature Communications, scientists have developed an Artificially Intelligent Lab Assistant (AILA) that can independently run complex laboratory experiments, analyse data, and make decisions in real time.

Until now, artificial intelligence tools such as ChatGPT were largely limited to assisting researchers with writing, calculations, or data analysis. IIT researchers on Tuesday shared that AILA goes several steps further. It can operate real scientific equipment, particularly the Atomic Force Microscope (AFM), one of the most delicate and complex instruments used to study materials at the nanoscale.

The research, titled ‘Evaluating large language model agents for automation of atomic force microscopy’ was led by IIT Delhi in collaboration with institutions in Denmark and Germany.

The first author, Indrajeet Mandal, a PhD scholar at IIT Delhi, said AILA has dramatically reduced the time needed for routine experimental tasks. Mandal said, “Earlier, it took an entire day to optimise microscope settings for clear images. Now, AILA completes the same task in just 7 to 10 minutes.”

AILA was developed under the supervision of Prof. N M Anoop Krishnan and Prof. Nitya Nand Gosvami.

Prof Krishnan explained that this marks a shift in how AI is used in science. “Earlier, AI could help you write about science. Now, it can actually do science designing experiments, running them on real equipment, and interpreting results,” added Prof Krishnan. Prof Gosvami highlighted the significance of the achievement, noting that mastering an Atomic Force Microscope typically takes years of training.

“AILA performing these tasks autonomously represents a paradigm shift in experimental research,” he said. The project involved researchers from IIT Delhi, Aalborg University in Denmark, institutions in Germany.