The Uprising of Machines is Getting Closer with Soft Skin and Wetware
Learn about the latest breakthroughs in artificial skin, touch sensors, and wetware computing using human brain organoids. The implications of conscious machines as bio-AI hybrids.
Hello once again, my dear readers, and welcome back to another futuristic episode on Tech Trendsetters, where we discuss technology, science, and the future trajectory of our world. Today, we talk about.. well, robots and how close we are to the human-like machines.
The rapid advancement of technology is bringing us closer to a future where machines possess human-like capabilities. Two areas of research that are pushing the boundaries of what's possible are artificial skin and touch sensors, and wetware computing using brain organoids.
Soft Skin-Like Touch Sensors may give Robots the Delicate Human Touch
Imagine a world where robots can feel and respond to touch with the same delicacy and precision as a human. Thanks to groundbreaking research by scientists from UBC and Honda Research Institute, this may soon become reality. Scientists have developed a soft capacitive sensor skin that can discriminate between normal force and shear, mimicking the properties of human skin.
Historically, one of the greatest challenges in robotics has been the lack of proper sensory feedback, especially when it comes to touch. Tasks that are simple for humans, like grasping a paper cup without crushing it, have been near impossible for robots. Without a sophisticated sense of touch, robots struggle with delicate manipulation and safe interaction in unstructured environments.
As you can guess, researchers has developed an innovative sensor skin that could change all that. The new sensor is composed of patterned elastomer and contains pillars that allow it to deform and buckle, much like real skin. By using the summation and differences of signals from four deformable capacitors, the sensor can differentiate between simultaneously applied normal force and shear with minimal cross-talk. It's even sensitive enough to detect finger proximity up to 15 mm away, allowing it to anticipate touch before contact is even made.
The key breakthroughs are the sensor's softness, smoothness, and ability to discriminate between types of touch with minimal cross-talk between sensing directions. It achieves an impressive normal stress sensitivity of 0.49 kPa and shear stress sensitivity of 0.31 kPa. That means it can detect the gentlest of touches.
All of this is achieved through a relatively simple fabrication method, making it a promising candidate for implementation in humanoid robotics applications.
Just think, robots of the future covered in this artificial sensing skin, able to delicately manipulate objects, safely interact with humans, and "feel" their environment in ways that open up entirely new possibilities. Robot surgeons performing the most intricate procedures. Robotic companions and caretakers for the elderly and disabled. The applications are endless.
Brain Organoids Offer a Revolutionary Platform for Wetware Computing Research
What if we could harness the computational power of human brain cells to create a revolutionary new form of biocomputer? Enter brain organoids and the audacious research into wetware computing and organoid intelligence.
Scientists have developed a first-of-its-kind open access "Neuroplatform" that enables researchers worldwide to remotely conduct experiments on lab-grown brain organoids. This bioprocessor is made by a Swiss company called SynEx. These 3D clusters of neurons are grown from human! induced pluripotent stem cells, not from rats or mice. The use of human cells is a deliberate choice in attempt create a more direct model of the human brain.
The brain organoids exhibit complex neural activity and offer an unprecedented platform to study how biological neural networks could be programmed to perform computational tasks. The Neuroplatform provides 24/7 automated maintenance of these organoids, which can live over 100 days in this system. That's a remarkable lifespan that allows for extended experimentation.
Researchers can remotely monitor the electrical activity of the neurons using a multi-electrode array and deliver precise electrical stimulations to influence the neural activity. But here's the really fascinating part – the platform also has microfluidic systems for delivering neurotransmitters like dopamine with pinpoint spatiotemporal accuracy.
Why is this so exciting? Well, dopamine is the brain's reward chemical. It's what makes us feel good when we accomplish something and it plays a crucial role in motivation and learning. By precisely delivering dopamine to the brain organoids when they exhibit desired activity patterns, the researchers are essentially training the neural networks using a reward system, just like how our brains learn.
This has huge implications not just for training wetware computers, but also for understanding how we, humans, can optimise our own brain power – dopamine reward is the key!
So while the technical details of microfluidics and spatiotemporal accuracy may sound complex, the key takeaway is this: by leveraging the brain's natural reward system, we could unlock new levels of intelligence in both artificial and natural neural networks. And that's an absolutely thrilling prospect.
The scale of research enabled by this platform is staggering.
Over 1,000 human brain organoids have already been cultured and studied, over 18 terabytes of electrophysiological data that is openly shared with the global research community. Experiments have demonstrated how induced global activity patterns can be used to store and process information across the 3D neural network. Closed-loop stimulation has been used to provide real-time feedback to the organoids, rewarding specific activity patterns to induce learning and plasticity.
While still in its early stages, this wetware computing research is yielding fascinating insights into the computational potential of biological neural networks. The energy efficiency is particularly amazing – the human brain cells is estimated to be thousands times more energy efficient than the best silicon-based AI systems today. Imagine a future where biocomputers grown from human neurons tackle the most complex machine learning problems while consuming a tiny fraction of the power.
Beyond the computing angle, the research also deepens our fundamental understanding of how the human brain develops and learns. It could unlock new approaches to brain-computer interfaces, artificial intelligence, allow in-vitro testing of how pharmaceuticals impact human neurology, and help us understand and treat neurological disorders in radically new ways. The potential upsides are immense.
Towards a Visionary Future
The breakthroughs in soft touch sensors and brain organoid computing foreshadow a future where the boundaries between biological and artificial begin to dissolve. Machines could soon possess the delicate touch of human skin and be powered by lab-grown human neurons.
This confluence of technologies opens up transformative possibilities but also raises profound concerns:
The reason for concern isn't necessarily just the ethics of the matter, but the fact that we barely understand AI as it is. Even the people who develop AI systems confirm that it's more akin to learning than programming in the sense of regular software. This means that if something goes wrong, the problematic code can't simply be removed.
Controlling AI is already a huge challenge, even without bionic components. As we've discussed the "superalignment” problem in previous episodes – addding biological components like brain organoids only compounds this challenge.
It's important to note that these brain organoids, despite their small size, already contain all the basic structures found in a regular human brain. Granted, without a body, they may not develop to their full potential, but that could simply be a matter of time. As the technology advances, we could see more sophisticated bio-AI hybrids that blur the line between natural and artificial cognition.
Brain organoids also have applications beyond robotics. They could be used in servers that power the internet, leveraging the incredible energy efficiency of biological neural networks. The human brain is already the most efficient computer on the planet in terms of energy consumption and computational density. Harnessing that efficiency in our computing systems could be a game-changer.
Perhaps the most profound question is whether we will create a conscious robot sooner or later, and what the ethical implications of that would be. How will we define and recognise machine consciousness? What rights would a conscious robot have? Are they allowed to breed? These are complex philosophical and legal questions for us as a society.
As we stand on the precipice of this bio-AI revolution, it's clear that we are entering uncharted territory. The potential benefits are immense – from more energy-efficient computing to advanced robotics to new insights into our own minds. But the risks and ethical challenges are equally profound. It will take ongoing research, public discourse, and proactive policy to ensure that we steer this technology towards beneficial outcomes for humanity.
One thing is certain for me – the future will be unlike anything we've seen before. The convergence of AI, robotics, and biotechnology is set to reshape our world in ways we can scarcely imagine.
So stay engaged, stay informed, and keep wrestling with these big questions, dear readers. The adventure is just beginning!
🔎 Explore more: