The complexity of biology has long been a double-edged sword for scientific and medical progress. On one hand, the intricacy of systems (like the human immune response) offers countless opportunities for breakthroughs in medicine and healthcare. On the other hand, that very complexity has often stymied researchers, leaving some of the most significant medical challenges—like cancer or autoimmune diseases—without clear solutions. The field needs a way to decipher this incredible complexity. Could the rise of agentic AI, artificial intelligence capable of autonomous decision-making and action, be the key to breaking through this impasse? Agentic AI is not just another tool in the scientific toolkit but a paradigm shift: by allowing autonomous systems to not only collect and process data but also to independently hypothesize, experiment, and even make decisions, agentic AI could fundamentally change how we approach biology. The mindboggling complexity of biological systems To understand why agentic AI holds so much promise, we first need to grapple with the scale of the challenge. Biological systems, particularly human ones, are incredibly complex—layered, dynamic, and interdependent. Take the immune system, for example. It simultaneously operates across 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 vast complexity. The problem lies in the sheer volume and interconnectedness of biological data. The immune system alone involves interactions between millions of cells, proteins, and signaling pathways, each influencing the other in real time. Making sense of this tangled web is almost insurmountable for human researchers. Enter AI agents: How can they help? This is where agentic AI steps in. Unlike traditional machine learning models, which require vast amounts of curated data and are typically designed to perform specific, narrow tasks, agentic AI systems can ingest unstructured and diverse datasets from multiple sources and can operate autonomously with a more generalist approach. Beyond this, AI agents are unbound by conventional scientific thinking. They can connect disparate domains and test seemingly improbable hypotheses that may reveal novel insights. What might initially appear as a counterintuitive series of experiments could help uncover hidden patterns or mechanisms, generating new knowledge that can form the foundation for breakthroughs in areas like drug discovery, immunology, or precision medicine. These experiments are executed at unprecedented speed and scale through robotic, fully automated laboratories, where AI agents conduct trials in a continuous, round-the-clock workflow. These labs, equipped with advanced automation technologies, can handle everything from ordering reagents, preparing biological samples, to conducting high-throughput screenings. In particular, the use of patient-derived organoids—3D miniaturized versions of organs and tissues—enables AI-driven experiments to more closely mimic the real-world conditions of human biology. This integration of agentic AI and robotic labs allows for large-scale exploration of complex biological systems, and has the potential to rapidly accelerate the pace of discovery. From agentic AI to AGI As agentic AI systems become more sophisticated, some researchers believe they could pave the way for artificial general intelligence (AGI) in biology. While AGI—machines with the capacity for general intelligence equivalent to humans—remains a distant goal in the broader AI community, biology may be one of the first fields to approach this threshold. Why? Because understanding biological systems demands exactly the kind of flexible, goal-directed thinking that defines AGI. Biology is full of uncertainty, dynamic systems, and open-ended problems. If we build AI that can autonomously navigate this space—making decisions, learning from failure, and proposing innovative solutions—we might be building AGI specifically tailored to the life sciences. Owkin’s next frontier: Unlocking the immune system with agentic AI Agentic AI has already begun pushing the boundaries of what’s possible in biology, but the next frontier lies in fully decoding one of the most complex and crucial systems in human health: the immune system. Owkin is building the foundations for an advanced form of intelligence—an AGI—capable of understanding the immune system in unprecedented detail. The next evolution of our AI ecosystem, called Owkin K, could redefine how we understand, detect, and treat immune-related diseases like cancer and immuno-inflammatory disorders. Owkin K envisions a coordinated community of specialized AI agents that can autonomously access and interpret comprehensive scientific literature, large-scale biomedical data, and tap into the power of Owkin’s discovery engines. These agents are capable of planning and executing experiments in fully automated, robotized wet labs, where patient-derived organoids simulate real-world human biology. The results of these experiments feed back into the system, enabling continuous learning and refinement of the AI agents’ models. What makes Owkin K particularly exciting is its potential to tackle the immune system—a biological network so complex that human intelligence alone has struggled to unravel it. By deploying AI agents with the ability to explore this intricate web autonomously, the project could reveal new therapeutic targets and strategies for immuno-oncology and autoimmune diseases, potentially accelerating the development of groundbreaking treatments. Navigating challenges and ethical considerations of agentic AI Of course, such powerful technology comes with significant challenges and ethical considerations, including trust, security, and transparency. But we must tackle these challenges as agentic AI becomes more integrated into healthcare and research. For example, we can develop mitigation plans that include rigorous validation protocols, real-time human oversight, and regulatory frameworks designed to ensure safety, accountability, and transparency. By prioritizing ethical design and close collaboration between AI systems and human experts, we can harness the potential of agentic AI while minimizing its risks. The future of biological research with agentic AI Agentic AI has the potential to reshape not just healthcare, but the very foundations of biological research. By allowing autonomous systems to explore the unknown, we may unlock new levels of understanding in areas like immunology, neuroscience, and genomics—fields that are currently constrained by the limits of human comprehension. We could soon see a world where AI-driven labs operate around the clock, pushing the boundaries of biology at speeds and scales that far exceed human capabilities. This would not only accelerate scientific discovery but also create new possibilities for personalized medicine, disease prevention, and even longevity. In the end, agentic AI may be more than just another tool for researchers. It could be the key to understanding life itself—one autonomous decision at a time. Davide Mantiero, PhD, Eric Durand, PhD, and Darius Meadon also contributed to this article. This content was produced by Owkin. It was not written by MIT Technology Review’s editorial staff.