The science of smell is entering a transformative era. In a recent perspective published in the Proceedings of the National Academy of Sciences (PNAS), Pritish Kumar Varadwaj explores how the convergence of human olfactory perception and artificial intelligence is opening entirely new dimensions in our understanding of smell. Rather than viewing olfaction merely as a sensory system that detects pleasant or unpleasant odors, the article argues that smell represents a rich information-processing network capable of revealing hidden aspects of biology, health, behavior, and environmental interactions.

Among human senses, olfaction remains one of the least understood. While enormous advances have been made in vision and hearing research, smell continues to present unique scientific challenges. Odors are generated by complex mixtures of volatile molecules, and their perception is shaped not only by chemical structure but also by genetics, physiology, memory, emotion, culture, and personal experience. This complexity makes olfaction both fascinating and difficult to model using conventional scientific approaches.

The emergence of artificial intelligence is beginning to change this landscape. AI systems can analyze vast amounts of chemical, biological, and sensory data simultaneously, uncovering patterns that may be impossible for humans to detect through traditional methods alone. By integrating computational approaches with experimental olfactory research, scientists are gaining new opportunities to decode relationships between molecular properties and perceived odor qualities. Such advances may accelerate our ability to predict smell, understand sensory diversity, and uncover previously hidden biological signals encoded within odor information.

Importantly, the significance of this approach extends far beyond fragrance science. Smell is increasingly recognized as a window into human health. Changes in body odor, breath composition, and environmental chemical signatures can reflect underlying physiological processes, including metabolic alterations, infectious diseases, neurological disorders, and aging. AI-enhanced olfactory analysis therefore has the potential to support non-invasive diagnostics, precision medicine, and early disease detection. The integration of machine intelligence with sensory science may enable researchers to identify subtle odor patterns that serve as biomarkers of health and disease long before symptoms become apparent.

Another exciting aspect is the possibility of bridging human and machine perception. Humans possess remarkable abilities to interpret odors in emotionally meaningful ways, while AI excels at detecting statistical relationships across complex datasets. Combining these complementary strengths could create a powerful framework for exploring how sensory information is encoded, interpreted, and translated into behavior. Such efforts may ultimately lead to smarter electronic noses, improved environmental monitoring technologies, personalized healthcare applications, and novel approaches to understanding the chemical world around us.

The perspective highlights a broader scientific vision: olfaction should not be regarded as an isolated sensory curiosity but as a multidimensional source of biological and environmental information. As artificial intelligence continues to advance, researchers may be able to access previously “unseen” dimensions of smell, transforming olfactory science into a highly predictive and data-driven discipline. The intersection of sensory neuroscience, computational biology, and AI therefore represents one of the most promising frontiers in modern science.

Ultimately, this work serves as a reminder that smell is far more than a mechanism for detecting odors. It is a gateway to understanding human biology, behavior, health, and our interaction with the environment. By combining the strengths of human perception and artificial intelligence, the next generation of olfactory research may unlock insights that were previously beyond our reach.

Reference: Pritish Kumar Varadwaj. Exploring the Unseen Dimensions of Olfaction via Human Perception and AI. Proceedings of the National Academy of Sciences (PNAS), 2026. PMID: 42709825.