Description

In a breakthrough for biocomputing, researchers at Indiana University Bloomington have created a system called 'Brainoware' that uses living human brain cells to perform speech recognition tasks. The system consists of brain organoids grown from stem cells and placed on top of a microelectrode array. After two days of training with 240 audio clips of eight people pronouncing Japanese vowel sounds, the organoids were able to recognize the voice of one individual with an accuracy of 70-80%. This adaptive learning process did not involve any feedback or supervision and relied on unsupervised learning. The researchers hope that this system could help overcome the high energy consumption and limitations of conventional AI, which often separates information and processing. However, critics argue that current deep-learning models are already better than any brain at specific tasks. Additionally, the organoids can only be maintained for one or two months, limiting their potential use in AI computing.