Scientists from various disciplines have presented a roadmap for the production of biocomputers from human brain cells. The initiative consists of creating bioinformatics systems using three-dimensional cultures of brain cells, called organodes, as biological material. They call this new interdisciplinary field “Organode Intelligence” (IO). The researchers’ goal is to develop IO as “a form of true biological computing utilizing brain organoids by leveraging advances in science and bioengineering in an ethically responsible manner.”
What is Organode Intelligence?
The human brain has inspired AI – which can perform a variety of tasks, from diagnosing diseases to creating intelligent content. However, the brain, the original model, continues to outperform AI in many ways: AI’s computing power pales in comparison to that of the human brain. For many researchers, working directly on the surface is therefore preferable to making the AI more like the brain. A group of scientists on Tuesday presented plans for a project to push computing in this direction. The roadmap describes what they call “Organode Intelligence” (IO).
We call this new interdisciplinary field organode intelligence. ‘A group of top scientists has come together to develop this technology, which we believe will usher in a new era of fast, powerful and efficient bioinformatics,’ said Professor Thomas Hartung of Johns Hopkins University. Published in the journal Frontiers in Science, the plan aims to develop biocomputers powered by human brain cells. Experts say these biocomputers could be faster, more efficient and more powerful than silicon-based computers and AI, while using a fraction of the power.
According to the new study, “organode intelligence” describes an emerging multidisciplinary field working on the development of biological informatics using three-dimensional cultures of human brain cells and brain-machine interface technologies, and which requires current brain organodes to be large, complex and durable three-dimensional structures enriched with cells and genes associated with learning. IO also involves connecting these brain organodes to next-generation input and output devices and artificial intelligence or machine learning systems.
For IO to be a successful field, one must learn new model, algorithm, and interface technologies for communicating with brain organodes, understand how they learn and compute, and process and store the vast amounts of data that brain organodes will generate.
Could brain organodes make good computers?
Brain organoids are cell cultures grown in the laboratory that share important aspects of brain function and structure, such as neurons and other brain cells essential for cognitive functions such as learning and memory. Brain organoids are not “mini brains”. While most cell structures are flat, brain organodes have a three-dimensional structure, which the researchers say increases cell density in culture by 1,000-fold. This means that neurons can make many more connections.
But even if brain organodes are good mimics of brains, why would they make good computers? According to Hartung, the brain is better at learning when silicon-based computers are better at dealing with numbers. For example, AlphaGo – the DeepMind AI that beat the world’s number one Go game in 2017 – was trained on data from 160,000 games. A person would have to play five hours a day for over 175 years to experience these many parts, explained Professor Hartung. The brain is a superior learner, but it’s also more energy efficient.
According to the study, the amount of energy expended to propel AlphaGo is greater than that required to sustain an active adult for a decade. Hartung said the brain also has an amazing capacity for storing information, estimated at around 2,500 trabytes. The professor explained that humans are reaching the physical limits of silicon computers because they can’t fit more transistors into a tiny chip (Moore’s Law). However, the brain is wired completely differently and has around 100 billion neurons connected by over 1015 connection points.
Hartung believes this is a huge performance difference compared to the world’s current technology. According to the study, OI research could give scientists a better understanding of how the human brain works, learning and memory, and potentially help find treatments for neurological conditions like dementia.
What Would Organode Intelligence Biocomputers Look Like?
According to Hartung, current brain organodes need to be upscaled for IO. They are too small and contain about 50,000 cells each. For the IO, we should bring that number to 10 million, he explained. At the same time, the researchers are also developing technologies to communicate with the organodes, i.e. to send them information and to read what they “think”. Researchers plan to adapt tools from different scientific disciplines, such as bioengineering and machine learning, and to design new stimulation and recording devices.
We have developed a brain-machine interface device, which is a kind of EEG cap for organodes, which we presented in a paper published last August. It’s a flexible shell densely covered with tiny electrodes that can pick up signals from and relay signals to the organode, Hartung explained. The authors envision that the OI will eventually integrate a wide range of stimulation and recording tools. These will orchestrate interactions through networks of interconnected organodes that implement more complex computations.
Researchers can create brain organoids from adult tissue using the revolutionary technique pioneered by Nobel Prize winners John Gurdon and Shinya Yamanaka.
Some Ethical Implications of Organode Intelligence
The authors point out in their article that it is important to take an integrated ethical approach to ensure that the Organod intelligence develops in a way that is compatible with ethics and society. According to the researchers’ report, some of the ethical concerns raised in the manufacture of human brain organoids are whether they can develop consciousness, even in rudimentary form, whether they can experience pain and suffering, and what rights people have, their cells are used to make brain organodes.
As part of this approach, Hartung et al state that interdisciplinary and representative teams of ethicists, researchers and members of the public will identify, discuss and analyze ethical issues. An important part of our vision is to develop IO in an ethical and socially responsible way. That’s why we teamed up with theologians early on to establish an “integrated ethics” approach. “Ethical issues are continually evaluated by different teams as research evolves,” Hartung said.
How far away are we from such a biocomputer?
Although IO is still in its infancy, a recent study by one of the paper’s co-authors – Dr. Brett Kagan from Cortical Labs – the proof of concept. His team has shown that a culture of normal, flat brain cells can learn to play the video game Pong. Her team is already testing this with brain organodes. And I would say that reproducing this experiment with Organodes already fits the basic definition of IO. From there, it’s just a matter of creating the community, tools, and technology needed to unlock IO’s full potential, Hartung concluded.
What are the reviews saying about Organode Intelligence?
The researchers’ paper aroused interest, but also questions about the feasibility of the project and ethical concerns. According to some critics, if the theories become reality, the difference in organode energy consumption is a major selling point. It only takes 12 watts to power a human brain, which is exceptionally efficient, especially when compared to the energy required for machine learning. If energy efficiency is an integral part of IO, it would be a big step forward and perhaps a viable platform for true general purpose artificial intelligence, they note.
But concepts like biocomputers and organod intelligence could give rise to a library of new ethical discussions. Conversations of organodes becoming sentient, sentient or self-aware and the resulting implications have been going on for years, although the technology is still immature. There is probably no technology without unintended consequences. It’s hard to rule out such risks, but as long as the human controls the input and output and the feedback to the brain about the consequences of their output, the human is in control, Hartung notes.
But as with AI, once we give AI/OI autonomy, the problem arises. Machines, whether silicon-based or cell-based, should not decide human life, he added. Hartung insisted that members of the research team with a background in medical ethics (biothics) endeavor to assess the ethical implications of working with the OI. Other critics also believe the technology is not viable. Their opinion is based on the idea that “effectively connecting human cells to machines is difficult and very expensive”.
Organode intelligence and biocomputers will not pose a threat to AI or old-fashioned human brains any time soon. But Hartung believes it’s time to ramp up production of brain organodes and train them in AI to fill some of the gaps in current silicon systems. It will be decades before we reach the goal of anything comparable to any type of computer. But if we don’t start creating funding programs for it, it will be much more difficult,” said Hartung.
Source: Organode Intelligence Study Report
And you ?
What is your opinion on the topic?
What do you think of the concept of organode intelligence?
What do you think of biocomputers? Do they represent the future?
In your opinion, is the technology presented in the study workable?
Do you think it could become a consumer technology? For what ?
Is Organode Intelligence a platform towards general artificial intelligence (AGI)?
See also
Game maker Valve is collaborating with OpenBCI to develop the next generation of direct-to-the-brain interface video games that deliver a fully immersive experience
Brain implant company backed by Bezos and Gates is reportedly testing mind-controlled computing systems on humans
Ultrasound can read monkey brains, opening up a new way of controlling machines through thought. Patients may prefer this less invasive technology to chips
A brain-machine interface enables a paralyzed person to use robotic arms to feed themselves, the participant was successful in 17 out of 20 trials, i.e. h85%