Researchers from Indiana University of Bloomington, the University of Florida, and the University of Cincinnati School of Medicine have made a groundbreaking advancement in AI hardware with the development of “Brainoware,” a human brain on a chip. This innovation, detailed in their paper “Brain Organoid Computing for Artificial Intelligence,” represents a significant leap in biocomputing.
The team cultivated specialized stem cells into neuron clusters, or organoids, each less than a nanometer wide. These organoids, connected to a circuit board via electrodes, allow machine-learning algorithms to interpret their responses. Brainoware boasts advancements in complexity, connectivity, neuroplasticity, and neurogenesis, with minimal energy consumption and rapid learning capabilities.
In a practical test, Brainoware achieved a 78% accuracy in a speech recognition task, identifying speakers based on the organoid’s neural activity in response to electrical stimulation. While less accurate than traditional AI systems and requiring resources like a CO2 incubator, Brainoware’s energy efficiency is notable compared to the 8 million watts used by current AI hardware.
The researchers envision “organoid intelligence” (OI) as the future of computing, powered by living human brain cells. This technology could have applications in studying neurological conditions and cognitive aspects, offering a new dimension to AI computations and learning.
Sources include: Analytics India Magazine