Computers have relied on silicon chips for decades, but traditional hardware is approaching its physical limits. To solve complex problems and reduce energy consumption, scientists are turning to a powerful existing processor: the human brain. This field is called Organoid Intelligence (OI). It involves fusing lab-grown brain tissue with electronic chips to create “biocomputers” that could eventually outperform traditional supercomputers.
Before understanding the computer aspect, it is helpful to understand the biological component. Brain organoids are tiny, three-dimensional cultures of tissue derived from human stem cells. They are often called “mini-brains” in popular media, though they are only about the size of a pen dot.
Researchers at institutions like Johns Hopkins University create these by taking skin samples, reprogramming them into stem cells, and coaxing them to grow into brain cells (neurons). Unlike flat cell cultures in a petri dish, organoids have a 3D structure. This allows them to develop functional connections, or synapses, that mimic the architecture of a real human brain.
Currently, a typical organoid contains roughly 50,000 cells. While this is small compared to the 86 billion neurons in a human brain, it is enough to process basic signals and form networks.
The concept of biological computing moved from theory to reality with a project known as “DishBrain.” In 2022, researchers at Cortical Labs in Australia successfully grew a layer of human neurons on top of a multi-electrode array. They then connected this biological system to a computer simulation of the vintage video game Pong.
The results were startling:
You might wonder why we need biological computers when our phones and laptops are already so fast. The answer lies in two areas: energy efficiency and learning capability.
Modern supercomputers are incredibly powerful, but they are energy vampires. For example, the Frontier supercomputer in Kentucky exceeded the computational speed of a single human brain in 2022. However, achieving that speed requires 21 megawatts of power.
In contrast, the human brain operates on roughly 20 watts. That is barely enough energy to power a dim lightbulb. If scientists can harness the efficiency of biological neural networks, they could create AI systems that require a fraction of the energy used by current data centers.
Silicon chips process information linearly. They solve problems one step at a time, just very quickly. Brains process information in parallel. They can handle incomplete data, recognize patterns, and adapt to new situations much better than standard code. This makes organoid intelligence ideal for tasks that stump traditional computers, such as:
Creating a biocomputer requires a specific interface. You cannot simply plug a USB cable into a cell culture. The interface relies on Microelectrode Arrays (MEAs).
These are tiny grids of sensors and stimulators. The organoid sits directly on top of the MEA.
Professor Thomas Hartung at Johns Hopkins University is leading the charge to scale this technology. He estimates that for OI to compete with standard computers, organoids need to scale up to around 10 million neurons.
There are significant hurdles to reaching this scale:
Fusing human brain tissue with machines raises obvious ethical questions. Could these organoids develop consciousness? Could they feel pain?
Currently, the organoids are too primitive to possess consciousness or feelings. However, as they grow larger and more complex, the line blurs. To address this, Hartung and his team proposed the “Baltimore Declaration.” This is a framework calling for embedded ethics in the research process. It ensures that bioethicists are involved at every step of development, rather than waiting until the technology is fully mature to ask difficult questions.
While a commercially available biological laptop is likely decades away, Organoid Intelligence has immediate applications in medicine:
Are biocomputers faster than regular computers? Not yet. Silicon computers are still much faster at simple math and logic. Biocomputers excel at learning, pattern recognition, and energy efficiency. The goal is to create hybrid systems that use silicon for math and biology for complex reasoning.
Where do the brain cells come from? They come from induced pluripotent stem cells. Scientists usually take a skin sample from a donor, reverse-engineer the cells into a stem-cell state, and then reprogram them to grow into neurons. No embryos are used in this specific process.
Is Organoid Intelligence the same as AI? No. Artificial Intelligence (AI) runs on code and silicon chips. Organoid Intelligence (OI) runs on actual biological tissue. However, the two fields may merge to create hybrid AI systems.
Can an organoid feel pain? Current organoids do not have the complex structure or sensory inputs required to feel pain or suffer. They are essentially biological processors. Ethical guidelines monitor this closely as the technology advances.