Forget Silicon: Scientists Are Building the Next Generation of Computing From Living Brain Cells
Lab-grown blobs of human neurons are playing video games, guiding robots through mazes, and hinting at a future where 'artificial' intelligence is made of flesh, not chips.

Key points
- Brain organoids, tiny lab-grown clusters of human neurons roughly the size of a chia seed, contain up to 5 million cells and produce electrical activity similar to a premature baby's brain waves.
- A Melbourne startup has used organoids to play Pong and Doom; researchers at UC San Diego are using them to steer robots through mazes.
- Johns Hopkins University researchers are building biocomputing systems, physical devices that blend living organoid tissue with electronic hardware.
- Organoids grown from the cells of autistic donors are helping scientists study how neural development differs, research that could reshape how the condition is understood and treated.
- No federal animal welfare law currently covers brain organoids, but bioethicists warn the rules will need updating if organoids keep growing in complexity.
Your skin can grow a brain. Not a full one, but a working cluster of several million neurons that fire electrical signals and generate brain waves. Scientists call these clusters human brain organoids, meaning miniature lab-grown organ models, and they've been making them for years.
Now they're putting them to work.
What are brain organoids actually doing?
Organoids are being used to play video games, guide robots and form the cores of early biocomputers, computers that use biological tissue instead of purely electronic parts. A Melbourne startup has had organoids playing Pong and Doom. At UC San Diego (UCSD), they're steering spider-shaped robots through mazes.
Developmental biologist Alysson Muotri runs the organoid programme at UCSD's Sanford Stem Cell Institute. His lab grows organoids in the tens of thousands, keeping them at body temperature for months. Each one is about the size of a bee's brain, with roughly 2.5 million live neurons inside a snot-coloured blob you could lose in a petri dish. The remaining cells are non-neural glial cells that act as scaffolding.
Muotri's organoids respond to electrical pulses, remember them and start to anticipate them. "Whatever environment you put them in, the first thing that they do is try to connect," he told Wired AI, which first reported this research in depth.
Making an organoid doesn't require surgery. A small skin sample, or blood or teeth, is enough. Lab technicians introduce special proteins to adult cells, reverting them to a blank-slate embryonic state. From there, those so-called induced pluripotent stem cells, cells that have been chemically reset so they can become almost any tissue type, can be coaxed into becoming neurons.
Muotri's personal motivation is autism. His 18-year-old son is autistic and needs round-the-clock care. By growing organoids from the cells of autistic donors, including his son, Muotri hopes to see exactly where neural development diverges from typical patterns. Earlier research relied heavily on mice, which made comparisons to human brain development difficult. Our 11 August story on AI mapping the gene network behind schizophrenia shows how much the field is leaning on biological data that simply wasn't available a decade ago.
Should anyone be worried about the ethics?
Right now, organoids occupy a legal grey zone. They're not people, and they're not animals, so no federal animal welfare rules apply. Sociologist John Evans, co-director of UCSD's Institute for Practical Ethics, puts it plainly: there's currently no permission needed to run experiments on them.
Bioethicists are watching carefully. If organoids keep growing in complexity, the rules will need rethinking. The trickier question is whether any organoid could ever be conscious. Most philosophers of mind argue consciousness requires a body and lived experience, neither of which a blob in a glass tube has. The question is no longer purely theoretical.
| Research group | What organoids are doing | Scale |
|---|---|---|
| UCSD (Muotri lab) | Steering robots, autism research, space radiation studies | Tens of thousands grown at once |
| Melbourne startup | Playing Pong and Doom | Lab-scale prototype |
| Johns Hopkins | Biocomputing chip prototypes | Early-stage hardware |
| Various labs | Drug testing, disease modelling | Widespread, routine |
What does this mean for ordinary people?
For most people, this research won't touch daily life for years. Its nearest-term impact is medical: faster drug testing and a deeper understanding of conditions like autism. The computing angle is much further out.
What it challenges is the assumption that intelligence, artificial or not, must be built from silicon. That's worth keeping in mind as headlines fixate on the latest chatbot.
Honest verdict: the most interesting AI story right now might not involve a computer at all.



