Organoid Intelligence - Revolutionary computers powered by human brain cells

Scientists are almost ready to unveil a revolutionary path for the advancement of computing. It's called "organoid intelligence" (OI), where lab-grown brain organoids would act as biological hardware.

Umetna inteligenca se že dolgo zgleduje po človeških možganih. Ta pristop se je izkazal za zelo uspešnega. AI se ponaša z impresivnimi dosežki – od diagnosticiranja zdravstvenih stanj do pisanja poezije. Kljub temu izvirni model še vedno v mnogih pogledih prekaša stroje. Zato lahko na primer »dokažemo svojo človečnost« s trivialnimi slikovnimi testi na spletu. Znanstveniki pa so namesto poskušanja simuliranja človeških možganov, prešli kar neposredno do vira.

Many disciplines are working to create revolutionary biocomputers where 3D cultures of brain cells, called brain organoids, serve as biological hardware. They published their plan to realize this vision in the journal Frontiers in Science.

»To novo interdisciplinarno področje imenujemo ‘organoidna inteligenca’ (OI),« je dejal profesor Thomas Hartung z Univerze Johns Hopkins. »Skupnost vrhunskih znanstvenikov se je zbrala, da bi razvili to tehnologijo, za katero verjamemo, da bo začela novo dobo hitrega, zmogljivega in učinkovitega bioračunalništva.«

What are brain organoids and why can they be such good computers?

Brain organoids are a special type of laboratory-grown cell culture. Although brain organoids are not classified as “mini-brains,” they do share key aspects of brain function and structure, such as neurons and other brain cells that are essential for the cognitive functions of learning and memory. Why are they so special? Most cell cultures have a flat structure, but organoids are three-dimensional, which increases their cell density by 1000 times. As a result, neurons can form many more connections.

But even if brain organoids are good imitations of the brain, why would they be good computers? After all, aren't computers smarter and faster than brains?

»Medtem ko so računalniki na osnovi silicija zagotovo boljši pri številkah, so možgani boljši pri učenju,” je pojasnil Hartung. »Na primer, AlphaGo [AI, ki je leta 2017 premagal prvega igralca igre Go na svetu] je bil usposobljen na podlagi podatkov iz 160.000 iger. Človek bi moral igrati pet ur na dan več kot 175 let, da bi odigral enako število iger.«

Možgani niso samo boljši učenci, so tudi energetsko učinkovitejši. Na primer, količina energije, porabljene za usposabljanje AlphaGo, je večja od tiste, ki je potrebna za “vzdrževanje” aktivne odrasle osebe celo desetletje.

"The brain also has an incredible capacity to store information, estimated at 2,500 TB," Hartung added. “We are reaching the physical limits of silicon computers because we cannot pack more transistors into a small chip. But the brain is built completely differently. It has about 100 billion neurons connected through more than 1015 connection points. That's a huge difference in power compared to our current technology.”

What would organoid intelligence biocomputers look like?

According to Hartung, current brain organoids need to be enlarged if they are to serve as organoid intelligence. “They are too small, each containing around 50,000 cells. We should increase this number to 10 million for the OI," he explained.

In parallel, researchers are developing technology to communicate with organoids. Simply put, they are developing technology to send information and read data from which they could tell what the organoids are "thinking". They will help each other with tools from various scientific disciplines and adapt them for their purposes, such as bioengineering and machine learning. This will allow them to develop new stimulation and recording devices.

"We have developed a device that will act as an interface between the brain and the computer. It is a kind of EEG cap for organoids, which we presented in an article published last August. It is a flexible shell that is densely covered with tiny electrodes that can pick up signals from the organoid and transmit them to it," said Hartung.

The team of researchers envisions OI eventually incorporating a wide variety of stimulation and recording tools. These will orchestrate interactions in networks of interconnected organoids that would perform more complex computational tasks.

Organoid intelligence could help prevent and treat neurological conditions

Potencial organoidne inteligence sega dlje od samega računalništva. Zdravstvo je področje, ki bi lahko imelo največ od nove tehnologije. Nobelova nagrajenca John Gurdon in Shinya Yamanaka sta razvila prebojno tehniko, s katero je izvedljivo možganske organoide izdelati iz tkiva odraslega človeka. To pomeni, da lahko znanstveniki razvijejo prilagojene možganske organoide iz kožnih vzorcev pacientov z nevralnimi motnjami, kot je Alzheimerjeva bolezen. Nato lahko testirajo, kako genetski dejavniki, zdravila in toksini vplivajo na nevrološka obolenja.

"Cognitive aspects of neurological conditions could also be studied with OI," Hartung said. “For example, we could compare memory formation in organoids obtained from healthy people and Alzheimer's patients and try to correct for relative deficits. OI could also be used to test whether certain substances, such as pesticides, cause memory or learning problems."

Ethics and moral principles

Creating human brain organoids that can learn, remember and interact with their environment raises complex ethical questions. For example, could we develop consciousness even in a basic form? Can they feel pain or suffering? And what rights would people have over brain organoids made from their own cells?

Raziskovalci se zelo dobro zavedajo, da njihovo delo poraja številna etična vprašanja. »Ključni del naše vizije je razvoj OI na etičen in družbeno odgovoren način,« je dejal Hartung. »Zaradi tega smo že od samega začetka sodelovali z etiki, da bi vzpostavili pristop ‘vgrajene etike’. Vsa etična vprašanja bodo nenehno ocenjevale ekipe, sestavljene iz znanstvenikov, etikov in javnosti.«

How far are we from the first organoid intelligence?

OI is still in its infancy. A recently published study by researcher Dr. But Brett Kagan of Cortical Labs proves that the concept works. His team demonstrated that a normal, flat cell culture can learn to play the video game Pong.

"His team is already testing this with brain organoids," Hartung confirms. “I would say that repeating this experiment with organoids already meets the basic definition of OI. From here, it's just a matter of building communities, tools and technologies to realize the full potential of organoid intelligence."