A few months ago, AIM interviewed Dr. Thomas Hartung from Johns Hopkins University, who is attempting to create intelligence in a petri dish. He stated that by combining a class organ with approximately 10 million neurons and high-performance computing hardware, which is about the size of a zebrafish brain, we can create a biocomputer that utilizes the decision-making capabilities of neurons to generate a new type of artificial intelligence. Recently, researchers from Indiana University Bloomington, the University of Florida, and the University of Cincinnati College of Medicine published a paper titled "Computing with Artificial Intelligence Brain Organoids." Now, we have a brain chip that takes research a step further. Researchers are continuously pushing the limits of these small, brain-like structures. This yet-to-be-peer-reviewed paper introduces Brainoware as the future of artificial intelligence hardware, aiming to replace artificial neural networks (ANNs). "The question we want to ask is whether we can utilize the computational power of biological neural networks in brain-like organs," said lead researcher Feng Guo to Tech Xplore. "This is just a proof-of-concept validation that we can do this work." Guo and his research team cultivated specialized clusters of stem cells, which then evolved into neurons, the basic building blocks of the brain. A normal brain consists of 86 billion neurons, each capable of forming connections with up to 10,000 other neurons. The class organ composed of neurons generated in Guo's lab has a diameter smaller than one nanometer. It is connected to a circuit board through a series of electrodes, allowing machine learning algorithms to interpret the organ's responses. The paper highlights that Brainoware's brain organoids have made advancements in complexity, connectivity, neural plasticity, and neurogenesis compared to existing 2D culture and neuro-morphic chips. All of these advancements are achieved with minimal energy consumption and rapid learning capabilities. Neural Networks in Organoids for AI This novel biohybrid computer combines "brain organoids" with traditional artificial intelligence, demonstrating the ability to achieve 78% accuracy in speech recognition tasks. In a live demonstration, the research team converted 240 recordings of eight Japanese speakers into electrical pulses. They then trained an artificial intelligence to identify the speakers based on the neural activity generated by the brain-like organ. Although Brainoware is not as accurate as traditional AI-powered computing systems in speech recognition and requires resources such as CO2 incubators to sustain the organoids, it represents an important step towards more advanced biocomputing systems in the future. A typical human brain only requires 20 watts of power, in stark contrast to the massive 8 million watts consumed by current AI hardware utilizing artificial neural networks (ANNs). The researchers of the paper state that while currently promising, silicon chips inspired by the brain still face challenges in simulating brain functions to achieve efficient AI computing. Brainoware utilizes living biological neural networks within 3D brain organoids, providing a potential solution to existing hardware limitations. This demonstrates the potential of integrating human biology into computing to enhance capabilities. Hartung and his team refer to this as "organoid intelligence" or OI, which could be the next step in computing. These systems would be powered by living human brain cells. The Future is Organoid Intelligence Real-world applications such as solving nonlinear equations highlight the potential of this technology to learn from training data by reshaping the neural connections of organoid neural networks (ONNs). Brainoware uses living human brain organoids as dynamic reservoirs for "unsupervised learning," transforming time-dependent inputs into spatiotemporal sequences for artificial intelligence computation. Through spatiotemporal electrical stimulation, Brainoware enhances its computational performance and demonstrates learning capabilities on a chip through synaptic plasticity. Generating and maintaining organoids face challenges such as heterogeneity, low throughput, and varying activity. The current integration of microelectrode arrays (MEAs) with Brainoware is limited, prompting the need for innovations such as brain-machine interfaces and soft electrodes to improve connectivity with AI hardware. With OI, we can study the cognitive aspects of the nervous system and put our brains to the test. For example, we can compare memory formation in organoids from healthy individuals with those suffering from Alzheimer's disease and attempt to repair any defects. Alternatively, we can experiment with substances like pesticides to determine if they cause learning or memory problems. Meanwhile, Elon Musk is implanting chips in the human brain through NeuraLink, while researchers plan to implant brains into chips.