Could brain cells in a dish help computers solve complex problems?
We know computers are faster than humans, but are they really more intelligent than us? Could sheer technology — wires, microchips, and transistors ever achieve the computing power of pure biology — organized bundles of nerve cells with thousands of dainty projections?
Probably not, according to science. That’s why a large group of researchers across the United States and Europe are commencing a new field of research called organoid intelligence. In a study in Frontiers in Science, they outline their plan to develop brain-like organoids — lab-grown clumps of nerve cells resembling tiny brains — that could one day become “biocomputers”.
The effort comes as a recognition of computing technology’s current limitations:
“A supercomputer finally surpassed the computational power of a single human brain in 2022. But it cost $600m and occupies 680 square meters,” lead scientist Thomas Hartung says. The human brain on the other hand, is thought to be a million times more power- and data-efficient, using only a tiny fraction of the machine’s energy and learning time.
What’s more, computing technology is breaking away from Moore’s law: the observation that transistor numbers on microchips tend to double every two years with the pace of innovation. The law “has held for 60 years,” Hartung says. “But soon we won’t be able to physically fit more transistors into a chip. A single neuron, on the other hand, can connect to up to 10,000 other neurons.”
Artificial intelligence poses its own obvious limitations. Despite the notorious success of bots like ChatGPT, AI is a long way from reliable problem-solving with changing variables — such as self-driving, or choosing every picture with a traffic light to confirm it’s not a robot. Some questions need not only ultra-fast calculations but also a uniquely human lens.
So what could solve our society’s increasingly complicated problems? According to Hartung, developing artificial brains in Petri dishes that could first mimic human intelligence, and later be linked to an AI with brain-machine interfaces.
If this sounds more science fiction than science, well that’s because, at the moment, it still is. Mimicking human intelligence in cell cultures needs vast technological innovations across biotechnology that could take decades at least to achieve.
Currently, organoids contain less than 100,000 cells — the size of a pencil’s tip and only half the brain of a fruit fly. So first, scientists need to shoot up organoid size and complexity.
“We set out to produce brain organoids with about 10 million neural cells,” the authors write.
For this many cells, however, they also need new technologies in synthetic blood vessels or similar perfusion systems — ways to deliver nutrients to cells and remove waste and byproducts from them.
Even if the organoid survives, making it functional will be tricky:
“We’re building tools that will enable us to communicate with the organoids,” Hartung says.
Even a tiny brain capable of intelligent functions wouldn’t do much without first receiving signals from its environment. So to introduce an input, researchers could connect them to retinal organoids: groups of cells that imitate the eye and respond to precise signals of light.
To record the organoid output, they are building a device that resembles a tiny EEG or electroencephalograph — a set of wearable electrodes used widely in the clinic and the lab to detect brain waves.
The researchers’ end goal? Create self-operational organoids so complex that they could help AI solve problems through brain-machine interfaces. For this, they also need dedicated algorithms that could translate the language of organic matter into digital information, and vice versa.
“My dream is to form a channel of communication between an artificial intelligence program and an [organoid intelligence] system that would allow the two to explore each other’s capabilities,” Hartung says. “My expectation is that we’ll see a lot of surprises.”
At this point you might rightly wonder: what about the enormous, ethical elephant in the room? Could this technology become sentient or self-conscious? And what if it does?
Hartung is aware of the many concerns and uncertainties: from the hard problem of consciousness to the simpler ‘whose cells are we going to use’.
To address these, the researchers will work with an ‘embedded ethics approach’, involving bioethicists throughout the development to monitor progress and ensure they spot any issues. They also plan continuous dialogues with the public to find out how people feel about the technology and steer their research accordingly.
Even if organoid intelligence doesn’t fulfill its radical biocomputer promise, along the way, it could still revolutionize how we study the brain. Organoids could offer unprecedented access to functions like learning and memory, or diseases like Alzheimer’s — questions obscured by superficial brain imaging, rudimentary cell cultures, or the need to sacrifice animals for more invasive studies.
“It’s a starting point,” Hartung says. “ I see this like sequencing the first genes of the human genome project: the enabling technology is in our hands, and we’re bound to learn a lot on the way.”
