As artificial intelligence researchers begin to confront the limitations of today’s dominant computing architectures, a radically different approach is gaining momentum: using living human brain cells as computational hardware. These experimental systems, often described as biocomputers, represent a convergence of neuroscience, biotechnology, and information science. Although still at an early stage, they have already demonstrated the ability to perform simple tasks such as playing the game Pong and carrying out basic speech recognition. The surge of interest in this field is being driven by three key forces: abundant venture capital flowing into AI-adjacent ideas, major advances in growing brain tissue outside the body, and rapid progress in brain–computer interfaces that blur the line between biological and digital systems.
The foundation of this technology rests on decades of neuroscience research. For nearly 50 years, scientists have cultured neurons on microelectrode arrays to observe how they fire and communicate. Early attempts at two-way communication between neurons and electronics in the early 2000s created the first primitive biohybrid systems, but progress stalled until the rise of brain organoids. In 2013, researchers demonstrated that stem cells could self-organize into three-dimensional brain-like structures. These organoids rapidly became valuable tools for drug testing and developmental biology, especially when paired with “organ-on-a-chip” technologies. Today, however, their neural activity remains rudimentary compared with the organized patterns of the human brain, and experts agree that current systems are far from conscious.
The modern phase of biocomputers accelerated in 2022 when Melbourne-based Cortical Labs showed that cultured neurons could learn to play Pong in a closed-loop system. While the experiment was technically impressive, public attention focused heavily on the authors’ use of the phrase “embodied sentience,” which many neuroscientists felt overstated the system’s capabilities. In response to both scientific and media confusion, researchers later introduced the term “organoid intelligence” to describe this emerging field. While catchy, the phrase risks implying equivalence with artificial intelligence, despite the vast gulf in capability.
Commercialization is now moving rapidly. Companies and research groups across the United States, Switzerland, China, and Australia are racing to develop scalable platforms built on biocomputers. Swiss firm FinalSpark already provides remote access to neural organoids, while Cortical Labs is preparing to ship a desktop biocomputer known as the CL1. Academic projects are also growing more ambitious, with proposals such as using organoid systems to predict environmental disasters like oil spills. Yet current performance remains limited: these systems exhibit only basic responsiveness and adaptation, not anything resembling true cognition.
More practical near-term applications appear in medicine and toxicology. Organoids are being explored as alternatives to animal testing, tools for studying epilepsy, and platforms for assessing how chemicals influence early brain development. These advances are incremental but scientifically credible.
Beyond the lab, biocomputers raise profound ethical and philosophical questions. What constitutes intelligence? Could a network of human cells ever deserve moral consideration? How should society regulate systems that incorporate living tissue into machines? As the technology matures, debates over consciousness, personhood, and the ethics of merging human biology with computing are likely to intensify, making governance as crucial as innovation itself.