Agent AI is not just another tool in the scientific toolbox, but a paradigm shift: by enabling autonomous systems not only to collect and process data, but also to independently hypothesize, experiment, and even make decisions, agent AI could fundamentally change our approach to biology.
The astonishing complexity of biological systems
To understand why agentic AI holds so much promise, we must first grapple with the scale of the challenge. Biological systems, especially human ones, are incredibly complex—layered, dynamic, and interdependent. Take the immune system for example. It acts simultaneously on multiple levels, from individual molecules to entire organs, adapting and responding to internal and external stimuli in real time.
Traditional research approaches, while powerful, struggle to account for this enormous complexity. The problem lies in the sheer volume and interconnectedness of biological data. The immune system itself involves interactions between millions of cells, proteins, and signaling pathways, each affecting the other in real time. Making sense of this intricate web is nearly insurmountable for human researchers.
Enter AI agents: How can they help?
This is where agent AI comes into play. Unlike traditional machine learning models, which require vast amounts of managed data and are typically designed to perform specific, narrow tasks, agent-based AI systems can receive unstructured and diverse data sets from multiple sources and can work autonomously with a more general approach.
Furthermore, AI agents are unfettered by conventional scientific thinking. They can connect different fields and test seemingly unlikely hypotheses that can reveal new insights. What might initially seem like a counterintuitive series of experiments could help reveal hidden patterns or mechanisms and generate new insights that can form the basis for breakthroughs in areas such as drug discovery, immunology or precision medicine.
These experiments are performed at unprecedented speed and scale in robotic, fully automated labs where AI agents conduct trials in a non-stop, non-stop workflow. Equipped with advanced automation technologies, these laboratories can handle everything from ordering reagents, preparing biological samples to performing high-throughput screening. In particular, the use of patient-derived organoids – 3D miniaturized versions of organs and tissues – allows AI-driven experiments to more closely mimic real-world conditions of human biology. This integration of agent-based AI and robotic laboratories enables large-scale exploration of complex biological systems and has the potential to rapidly accelerate the pace of discovery.
From agent AI to AGI
As agent-based artificial intelligence systems become increasingly sophisticated, some researchers believe they could pave the way for artificial general intelligence (AGI) in biology. While AGI—machines with human-equivalent general intelligence capacity—remains a distant goal in the broader artificial intelligence community, biology may be one of the first fields to approach that threshold.
Why? Because understanding biological systems requires exactly the kind of flexible, goal-oriented thinking that defines AGI. Biology is full of uncertainties, dynamic systems and open problems. If we build an AI that can navigate this space autonomously—making decisions, learning from failures, and proposing innovative solutions—we could be building AGIs specifically tailored to the life sciences.